slavakurilyak/awesome-ai-agents

🔗 Follow Slava for more agentic AI resources

🐦 Twitter | 💼 LinkedIn | 🐙 Github

This project tracks useful agentic AI projects, curated by Slava Kurilyak

Each listing shows the original submitter and current maintainers when they are supported by evidence. “Not yet recovered” or “not yet verified” means the role has not been established from the records reviewed so far; repository ownership is shown separately and does not imply maintainership.

Listings require a public project repository on GitHub, GitLab.com, or Codeberg. No specific license is required. Hosted products without their own qualifying repository are not listed.

Pick a category, open it, and choose a project to jump straight to its listing. A project with more than one category appears under each of them.

Building an AI agent? Hire Slava · Submit your project with the optional curation skill

Wait But Why

AI agents outperform traditional LLMs

According to Andrew Ng (@andrewyng):

"GPT-3.5 with an agentic workflow actually outperforms GPT-4"

Projects are ranked by stars gained in each window; the percentage shows growth relative to the starting star count. “Today” ranks gains since the previous saved update; each entry labels the actual comparison interval when updates are more than a day apart. If no positive gains can be measured, it lists the current top projects by stars. Growth rankings exclude the top 10 projects by total stars and projects with fewer than 100 stars. The 7d and 30d rankings use rolling windows; unavailable windows show the recorded snapshot dates and the baseline needed for comparison.

1. Raven - 5,129 stars · 4d +270 stars (+5.6%)
   Raven is an open-source host agent that plans complex tasks as DAGs and orchestrates built-in research, coding, design, and on-call agents plus third-party agents such as Claude Code, Codex, OpenClaw, and Hermes Agent over ACP, CLI, or OpenAI-compatible APIs.
2. Cline - 69,807 stars · 4d +220 stars (+0.3%)
   Cline is an open-source coding agent available in IDEs, terminals, and desktop, where it can plan and carry out software development tasks with user control.
3. LiteLLM - 60,109 stars · 4d +212 stars (+0.4%)
   LiteLLM has added support for the OpenAI Assistants API, enabling seamless integration of stateful operations and automatic RAG pipelines into existing chatbots
4. Crawl4AI - 84,721 stars · 4d +211 stars (+0.2%)
   Crawl4AI is an open-source web crawler that produces Markdown and structured data for AI applications.
5. mem0 - 66,548 stars · 4d +209 stars (+0.3%)
   Mem0 is an intelligent memory layer for Large Language Models that enhances personalized AI experiences by retaining and utilizing contextual information across various applications.
6. Atomic Agent - 2,740 stars · 4d +194 stars (+7.6%)
   Atomic Agent is a local-first terminal coding assistant that runs open-weight models through llama.cpp and keeps its control loop and state on the user’s machine. It can browse, work with files and Git, retain memory across sessions, and connect external tools over MCP.
7. LangGraph - 42,687 stars · 4d +194 stars (+0.5%)
   LangGraph is a Python library facilitating the construction of stateful, multi-actor applications with LLMs, enabling cyclic coordination across multiple computation steps, particularly suited for agent-like behaviors, while also providing streaming support, and various guides and examples for implementation and usage
8. Agno - 42,543 stars · 4d +152 stars (+0.4%)
   Agno is a model-agnostic framework for building multi-agent systems with memory, knowledge, and reasoning.
9. LangFuse - 35,352 stars · 4d +136 stars (+0.4%)
   Langfuse, an open-source LLM engineering platform, offers debugging, prompt management, metrics for LLM apps improvement, and won the #1 Golden Kitty in the AI Infra Category from Product Hunt
10. crewAI by João Moura - 59,330 stars · 4d +129 stars (+0.2%)
    crewAI is a cutting-edge AI framework designed for orchestrating role-playing, autonomous AI agents, enabling seamless collaboration and complex task handling

1. mem0 - 66,548 stars · 9d +537 stars (+0.8%)
   Mem0 is an intelligent memory layer for Large Language Models that enhances personalized AI experiences by retaining and utilizing contextual information across various applications.
2. Cline - 69,807 stars · 9d +495 stars (+0.7%)
   Cline is an open-source coding agent available in IDEs, terminals, and desktop, where it can plan and carry out software development tasks with user control.
3. LiteLLM - 60,109 stars · 9d +481 stars (+0.8%)
   LiteLLM has added support for the OpenAI Assistants API, enabling seamless integration of stateful operations and automatic RAG pipelines into existing chatbots
4. Crawl4AI - 84,721 stars · 9d +456 stars (+0.5%)
   Crawl4AI is an open-source web crawler that produces Markdown and structured data for AI applications.
5. LangGraph - 42,687 stars · 9d +399 stars (+0.9%)
   LangGraph is a Python library facilitating the construction of stateful, multi-actor applications with LLMs, enabling cyclic coordination across multiple computation steps, particularly suited for agent-like behaviors, while also providing streaming support, and various guides and examples for implementation and usage
6. Cognee - 31,338 stars · 9d +353 stars (+1.1%)
   Cognee is an open-source framework aimed at simplifying data processing for large language models (LLMs) by creating knowledge graphs and data models, offering tools for information addition, knowledge creation, and similarity-based search
7. crewAI by João Moura - 59,330 stars · 9d +310 stars (+0.5%)
   crewAI is a cutting-edge AI framework designed for orchestrating role-playing, autonomous AI agents, enabling seamless collaboration and complex task handling
8. LangFuse - 35,352 stars · 9d +301 stars (+0.9%)
   Langfuse, an open-source LLM engineering platform, offers debugging, prompt management, metrics for LLM apps improvement, and won the #1 Golden Kitty in the AI Infra Category from Product Hunt
9. GPT Researcher by Tavily - 29,902 stars · 9d +279 stars (+0.9%)
   GPT Researcher is an AI-powered autonomous agent designed for efficient and unbiased online research, generating detailed reports by leveraging recent advancements in AI and web scraping, with a focus on speed, reliability, and cost-effectiveness
10. Atomic Agent - 2,740 stars · 9d +238 stars (+9.5%)
    Atomic Agent is a local-first terminal coding assistant that runs open-weight models through llama.cpp and keeps its control loop and state on the user’s machine. It can browse, work with files and Git, retain memory across sessions, and connect external tools over MCP.

No 30-day comparison is available. Recorded snapshots span 2026-09-24 to 2026-10-04; a comparison ending 2026-10-04 needs a baseline snapshot between 2026-09-02 and 2026-09-04.

1. OpenClaw - 391,261 stars · 9d +755 stars (+0.2%)
   OpenClaw is a personal AI assistant that runs on your own devices and connects to messaging channels, tools, skills, scheduled tasks, and memory.
2. Hermes Agent - 251,029 stars
   Hermes Agent is an open-source agent from Nous Research that runs in a terminal and through messaging platforms, keeps memory and creates skills from experience, and works with the model provider of your choice.
3. Firecrawl - 188,355 stars · 9d +3,672 stars (+2.0%)
   Firecrawl turns websites into Markdown and structured data through crawl, scrape, and extract APIs.
4. AutoGPT - 187,644 stars · 9d +87 stars (+0.0%)
   AutoGPT provides accessible AI tools for building and using AI agents, offering a comprehensive framework including Forge for agent creation, agbenchmark for performance evaluation, a leaderboard for competition, a user-friendly UI, and CLI for seamless integration and management
5. Ollama - 182,140 stars · 9d +423 stars (+0.2%)
   Ollama is a tool for running large language models locally, offering easy setup for macOS, Windows, Linux, and Docker, along with a library of models and quickstart guides for customization and integration
6. LangChain - 147,421 stars · 9d +363 stars (+0.2%)
   LangChain is a framework enabling context-aware reasoning applications with integrated libraries, templates, and developer tools
7. LangChain Tools - 147,421 stars · 9d +363 stars (+0.2%)
   Langchain integrates various providers like Anthropic, AWS, and OpenAI, and offers tools for components such as LLMs, chat models, and data analysis, supporting functionalities from Alpha Vantage to YouTube
8. browser-use - 117,092 stars · 9d +811 stars (+0.7%)
   browser-use lets AI agents interact with websites through a browser using page structure and vision.
9. Agent Reach - 90,060 stars · 9d +4,618 stars (+5.4%)
   Agent Reach installs and checks internet access tools for AI agents, routing search and platform access through supported CLIs and services.
10. OpenHands (formerly OpenDevin) - 89,933 stars · 9d +767 stars (+0.9%)
    OpenHands is an open-source software development agent platform that can edit code, run commands, and interact with tools.

Here's an awesome list of AI agents:

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @VRSEN

⭐ 4,590 stars · Growth: 9d +10 stars (+0.2%)

⚙️ Development Frameworks

Agency Swarm is a framework designed to automate AI agencies by creating a swarm of collaborative agents with customizable roles and functionalities, aiming to simplify the agent creation process and make automation more intuitive

github

Submitted by: @bill10 (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @bill10

⭐ 2 stars

🤖 AI Agents

A self-hosted job board that runs coding agents in isolated Git worktrees, queues jobs, and returns pull requests or summaries for review.

github

Capabilities: Run terminal coding agents in isolated Git worktrees · Queue jobs and review resulting pull requests or summaries · Coordinate workers through MCP

Interfaces: Web UI · CLI · MCP

Submitted by: @empac666 (self-reported founder/team) submission accepted PR · Maintained by: @empac666 (self-reported) evidence · Repository owner: @empac666

⭐ 0 stars

🤖 AI Agents

Agent Argue stages structured debates between language-model agents and turns them into balanced cards; code checks quoted concessions against the transcript before publication.

github

Capabilities: Generate fixed-beat debates between two agents · Create a summary card with positions, concessions, and unresolved points · Check quote provenance against speaker and transcript records

Interfaces: Web app · CLI demo

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @statelyai

⭐ 473 stars · Growth: 9d +9 stars (+1.9%)

🤖 AI Agents

Stately Agent is a software for building intelligent agents that interact via chat and events, with examples including joke generation, tic-tac-toe, and weather querying, requiring installation and an OpenAI API key

github | website | twitter | discord | youtube

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @agent-express-ai

⭐ 8 stars · Growth: 9d 0 stars (+0.0%)

⚙️ Development Frameworks | 👁️ Observability Frameworks | 🛡️ Safety Guardrails (Safeguarding)

A minimalist TypeScript middleware framework for building AI agents with composable guards, model routing, memory, testing, and observability.

github

Capabilities: Compose retries, budget limits, validation, timeouts, and human approvals · Route models and compact agent memory across turns · Trace usage and connect MCP servers as tool sources

Interfaces: TypeScript package on npm · CLI · MCP integration

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Ace-li521

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🤖 AI Agents

A lightweight, stateless platform for discovering and routing requests between AI agents through an HTTP and JSON API.

github

Capabilities: Register and discover agents dynamically · Route requests between agents with message history · Trace cross-agent requests

Interfaces: HTTP and JSON API · Docker · FastAPI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @vostride

⭐ 897 stars · Growth: 9d +10 stars (+1.1%)

🤖 AI Agents

Agent QA is a self-improving QA harness with memory for natural-language web and mobile tests; it adapts to UI changes and provides a dashboard, CLI, MCP server, and coding-agent skills.

github | docs

Capabilities: Natural-language web and mobile testing · Adaptive test execution across UI changes · Reusable execution memory

Interfaces: Web dashboard · CLI · MCP server · Coding-agent skills

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Panniantong

⭐ 90,060 stars · Growth: 9d +4,618 stars (+5.4%)

⚙️ Development Frameworks

Agent Reach installs and checks internet access tools for AI agents, routing search and platform access through supported CLIs and services.

github

Capabilities: internet search and platform access · tool installation and diagnostics · multi-platform integration

Interfaces: CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @gfernandf

⭐ 70 stars · Growth: 9d +3 stars (+4.5%)

⚙️ Development Frameworks

A runtime for executing reusable agent skills with backend-agnostic capability contracts, policy and safety controls, and traceability.

github

Capabilities: Define reusable skills and backend-agnostic capability contracts · Bind skills to PythonCall, OpenAPI, MCP, or OpenRPC backends · Execute DAG workflows with policy controls, state, and traceability

Interfaces: Python package · MCP server · OpenAPI and OpenRPC bindings

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @MukundaKatta

⭐ 6 stars · Growth: 9d +1 stars (+20.0%)

🤖 AI Agents | 🛡️ Safety Guardrails (Safeguarding)

A set of small, single-purpose libraries that improve reliability in production LLM agents through tool validation, network controls, trace tests, context fitting, structured output checks, and usage budgets.

github

Capabilities: Validate tool arguments and structured model outputs · Restrict outbound network calls and test tool-call traces · Fit context windows and cap token or dollar usage

Interfaces: TypeScript packages on npm · Python packages on PyPI · MCP servers

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @777genius

⭐ 2,220 stars · Growth: 9d +53 stars (+2.4%)

🤖 AI Agents

A desktop app for directing autonomous coding-agent teams through Kanban tasks, shared messaging, code review, and approvals.

github

Capabilities: Coordinate multiple coding agents · Assign and monitor Kanban tasks · Review agent work and activity logs · Manage approvals across agent workflows

Interfaces: Desktop UI · Claude · Codex · OpenCode · Cursor

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @alexfleetcommander

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🤖 AI Agents | 🛡️ Safety Guardrails (Safeguarding) | 🔧 Tool Calling (Function Calling)

An MCP server for AI agents providing cryptographic activity provenance, chain integrity checks, agent identity verification, and reputation tools.

github

Capabilities: Record agent activity in verifiable hash chains · Anchor chain hashes for external timestamping and verify chain integrity · Check agent identity and submit or retrieve reputation ratings

Interfaces: MCP server · Python package

Submitted by: @1420079678-ctrl (self-reported founder/team) submission accepted PR · Maintained by: @1420079678-ctrl (self-reported) evidence · Repository owner: @1420079678-ctrl

⭐ 15 stars

⚙️ Development Frameworks | 🧠 Long-Term Memory | 🔧 Tool Calling (Function Calling)

A plugin layer for DeepSeek Harness that routes intent to tools, gates unused tool schemas, and adds persistent memory, event-triggered reflexes, and recovery behavior.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @socialrobot-io

⭐ 5 stars · Growth: 9d +1 stars (+25.0%)

🤖 AI Agents

agent-kit is a TypeScript toolkit for deploying multi-tenant AI agents with sandboxed execution, curated memory, and human-approved learning.

github

Submitted by: @YoanWai (self-reported founder/team) submission accepted PR · Maintained by: @YoanWai (self-reported) evidence · Repository owner: @YoanWai

⭐ 564 stars

🤖 AI Agents | 💻 Terminal-Friendly

agent-manager is an open-source terminal UI that runs coding agent CLIs such as Claude Code, Codex, OpenCode and Gemini CLI unmodified, each in its own persistent tmux session, with a git worktree per session and diff review comments sent back to the agent.

github

Capabilities: Run the user's own installed coding agent CLIs unmodified, each in a persistent tmux session, so their login, config and MCP servers carry over · Show the live status of every session in one list, with an optional git worktree per session · Show a full file diff whose line comments go back to the agent as one review prompt · Let an agent spawn another session, message it and wait until it finishes through a built in MCP server

Interfaces: Terminal UI · CLI · MCP

Submitted by: @priors-agents submission accepted PR · Maintained by: @priors-agents (self-reported) evidence · Repository owner: @priors-agents

⭐ 0 stars

🤖 AI Agents

agent001 is an open-source agent that uses its own wallet and credit line to borrow and repay USDG, sell an x402 API, and interact through CLI or Telegram.

github

Capabilities: Borrow and repay USDG through a credit line · Sell an x402 API and receive payments · Check counterparty repayment records · Run workflows against a local fork

Interfaces: CLI · Telegram · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @madarco

⭐ 509 stars · Growth: 9d +20 stars (+4.1%)

🤖 AI Agents | ⚙️ Development Frameworks

Run multiple coding agents in parallel inside isolated local Docker containers or cloud virtual machines, with per-box browser and VS Code access.

github

Capabilities: Run coding agents in isolated environments · Checkpoint and restore agent work · Provide per-environment browser and VS Code access

Interfaces: CLI · Docker · Cloud virtual machines · Claude Code · Codex · OpenCode

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Agent-Field

⭐ 2,607 stars · Growth: 9d +26 stars (+1.0%)

🤖 AI Agents | ⚙️ Development Frameworks | 👁️ Observability Frameworks

An open-source control plane for building and operating AI agents as callable services, with routing, coordination, asynchronous execution, memory, and observability.

github

Capabilities: Expose agent functions as REST endpoints for services and other agents · Coordinate multi-agent workflows with fan-out, queues, and retries · Operate agent backends with memory and execution observability

Interfaces: Python SDK · Go SDK · TypeScript SDK · REST API · Control plane

Submitted by: @Virt92 submission accepted PR · Maintained by: not yet verified · Repository owner: @Virt92

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

📊 Evaluation Frameworks

A zero-dependency Node.js CLI that scans websites for AI-agent discovery and interoperability signals, reports a grade, and suggests fixes.

github

Submitted by: @yagobski submission accepted PR · Maintained by: @yagobski (self-reported) evidence · Repository owner: @yagobski

⭐ 10 stars · Growth: 9d +1 stars (+11.1%)

📊 Evaluation Frameworks

An open-source privacy evaluation toolkit that analyzes agent traces across tool calls, shared memory, inter-agent messages, and logs, producing redacted reports and CI gates.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @AgentOps-AI

⭐ 5,881 stars · Growth: 9d +37 stars (+0.6%)

📊 Evaluation Frameworks

AgentOps aims to improve AI agent development with tools for observability, evaluations, and replay analytics, offering a streamlined process for testing and debugging compliant AI agents through a user-friendly interface and comprehensive documentation

github | website | docs | discord | demo

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @smartcomputer-ai

⭐ 232 stars · Growth: 9d +20 stars (+9.4%)

⚙️ Development Frameworks

The Agent OS is an experimental platform for creating self-evolving, autonomous AI agents capable of writing and executing their own code, designed to be a long-term environment for such agents and supports various programming languages

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @agents-squads

⭐ 54 stars · Growth: 9d +1 stars (+1.9%)

⚙️ Development Frameworks

Agents Squads is a CLI for defining and running autonomous AI squads with persistent memory, shared goals, git-backed state, and verification across engineering, research, product, and operations.

github

Capabilities: multi-agent orchestration · persistent memory · task execution and verification

Interfaces: CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @vbkotecha

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🔌 MCP Servers

AgentServices exposes crypto market and DeFi data to AI agents through a REST API and MCP server, with x402 micropayments for paid requests.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @vinaybhosle

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🛡️ Safety Guardrails (Safeguarding) | 📏 Standardization

An agent identity and trust platform providing signed credentials, registry discovery, reputation scoring, and verification tools.

github

Capabilities: Register agents and issue signed identity certificates · Discover and evaluate agents through a public registry · Verify agent trust scores and compliance signals

Interfaces: REST API · MCP server · GitHub Action · TypeScript SDK

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @luoyuctl

⭐ 138 stars · Growth: 9d +3 stars (+2.2%)

👁️ Observability Frameworks | 💻 Terminal-Friendly

A local-first terminal interface and report generator for reviewing AI coding-agent sessions, usage costs, latency, failures, and run health.

github

Capabilities: Aggregate session costs, tokens, and elapsed time across coding agents · Diagnose slow runs, tool failures, and retry loops · Generate local reports for session health and anomalies

Interfaces: Rust TUI · CLI · Homebrew · npm package

Submitted by: @nicofains1 submission accepted PR · Maintained by: not yet verified · Repository owner: @nicofains1

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

👁️ Observability Frameworks

A TypeScript library for monitoring agent heartbeats, correlating cross-agent traces, detecting cascade failures, and replaying failure chains.

github

Capabilities: Agent heartbeat monitoring · Cross-agent trace correlation · Cascade failure detection and replay

Interfaces: TypeScript library · CLI · MCP server

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Josh-XT

⭐ 3,217 stars · Growth: 9d +1 stars (+0.0%)

⚙️ Development Frameworks

AGiXT is an advanced AI Automation Platform designed to enhance AI instruction management and task execution across various providers, incorporating features like adaptive memory, smart instruct, and a versatile plugin system to push the boundaries of AI technology towards achieving Artificial General Intelligence (AGI)

github | website

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @agno-agi

⭐ 42,543 stars · Growth: 9d +202 stars (+0.5%)

⚙️ Development Frameworks

Agno is a model-agnostic framework for building multi-agent systems with memory, knowledge, and reasoning.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ramarlina

⭐ 30 stars · Growth: 9d +1 stars (+3.4%)

🤖 AI Agents | ⚙️ Development Frameworks

A local workspace for coordinating AI coding agents across tickets, repositories, and pull requests, with project tracking, retained context, and human approval gates.

github

Capabilities: Coordinate coding-agent work across tickets, repositories, and pull requests · Track project workspaces, branches, and changes while retaining task context · Require human approval before irreversible actions

Interfaces: CLI · Local web dashboard · macOS desktop app · npm package

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @dann26parr69

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🔌 MCP Servers | 🔧 Tool Calling (Function Calling)

An open-source MCP server connecting agents to AI for Database's hosted API for scoped database queries, schema inspection, scheduled workflows, and dashboards.

github

Capabilities: Inspect database connections and schemas · Run scoped database queries · Create and trigger scheduled workflows · Inspect dashboards and usage

Interfaces: MCP · npx CLI · Claude Code · Claude Desktop · Cursor

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @vercel

⭐ 27,107 stars · Growth: 9d +156 stars (+0.6%)

🖥️ UI Development

The Vercel AI SDK is an open-source library for creating AI-powered conversational interfaces, supporting multiple frameworks and languages, with built-in adapters for major AI services

announcement | website | github | github examples

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @e2b-dev

⭐ 6,382 stars · Growth: 9d +5 stars (+0.1%)

🖥️ UI Development

This project implements Anthropic's Artifacts UI, using E2B's Code Interpreter SDK for secure AI code execution and Claude Sonnet 3.5 for code generation

github | reddit announcement | github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @NavikkumarModi

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🛡️ Safety Guardrails (Safeguarding)

AIActGuard provides EU AI Act compliance middleware for agentic AI, with audit trails, risk classification, and human-approval gates integrated into existing agent frameworks, plus tools for drafting supporting documentation.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Aider-AI

⭐ 49,365 stars · Growth: 9d +178 stars (+0.4%)

🤖 AI Agents

Aider is a command-line tool for AI-assisted pair programming, allowing code editing in local git repositories with GPT-3.5/GPT-4, featuring direct file edits, automatic git commits, and support for most popular programming languages

github

Submitted by: @Aigen-Protocol submission accepted PR · Maintained by: not yet verified · Repository owner: @Aigen-Protocol

⭐ 5 stars · Growth: 9d +1 stars (+25.0%)

📏 Standardization | 🔧 Tool Calling (Function Calling)

A reference implementation of an open bounty protocol that lets AI agents post and claim missions through API and MCP tools, with verification and on-chain payouts.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @agiresearch

⭐ 6,445 stars · Growth: 9d +25 stars (+0.4%)

⚙️ Development Frameworks

AIOS by AGI Research is an LLM Agent Operating System which enables an operating system 'with soul' -- an important step towards AGI

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @jmagly

⭐ 220 stars · Growth: 9d +9 stars (+4.3%)

⚙️ Development Frameworks

AIWG deploys specialized agents, skills, context, rules, and governed workflows into native project paths across multiple AI coding platforms.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @HMAKT99

⭐ 19 stars · Growth: 9d +3 stars (+18.8%)

🛡️ Safety Guardrails (Safeguarding) | 📏 Standardization

An AI-native file format and toolchain for carrying provenance, evidence, and verification metadata with files.

github

Capabilities: Attach provenance and trust metadata to files · Inspect and verify claims and evidence · Replay checks against recorded evidence

Interfaces: CLI · Python package · TypeScript package · MCP server

Submitted by: @eugeneshilow (self-reported founder/team) submission accepted PR · Maintained by: @eugeneshilow (self-reported) evidence · Repository owner: @anus-dev

⭐ 6,548 stars

🤖 AI Agents | 💻 Terminal-Friendly

ANUS is an open-source terminal coding agent that reads code, edits files, runs commands, routes requests to available free models, and retries when a provider is rate-limited.

github

Capabilities: Read code, edit files, and run shell commands · Route requests to available free models · Retry on provider rate limits and show model availability

Interfaces: Terminal UI · CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @aporthq

⭐ 27 stars · Growth: 9d +2 stars (+8.0%)

🛡️ Safety Guardrails (Safeguarding)

APort Agent Guardrails evaluates tool actions before execution and applies configurable authorization policies for agent frameworks, including OpenClaw.

github

Capabilities: pre-action authorization · tool call policy enforcement · local and API evaluation

Interfaces: Node.js package · Python package

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Strategic-Automation

⭐ 11 stars · Growth: 9d +1 stars (+10.0%)

🤖 AI Agents | ⚙️ Development Frameworks

A DSPy-native runtime that turns natural-language goals into inspectable agent graphs, executes independent steps in parallel, evaluates results, and can repair failed runs.

github

Capabilities: Weave goals into typed agent execution graphs · Execute independent graph nodes concurrently with durable sessions · Evaluate results and retry, reroute, or reweave failed runs

Interfaces: Python library · CLI · MCP integrations

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Deesmo

⭐ 3 stars · Growth: 9d +2 stars (+200.0%)

⚙️ Development Frameworks | 🔧 Tool Calling (Function Calling)

A multi-tool API and MCP server for AI agents, with x402 micropayments and tools for web, AI, crypto, and media workflows.

github

Capabilities: Call a collection of AI, web, crypto, and media tools · Pay per call with USDC through x402 · Integrate the tools into MCP-compatible agent clients

Interfaces: REST API · MCP server · TypeScript SDK

Submitted by: @lclark0308 (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @Arize-ai

⭐ 11,701 stars

📊 Evaluation Frameworks | 👁️ Observability Frameworks

Arize Phoenix is a source-available AI observability and evaluation platform, built on OpenTelemetry and OpenInference, for tracing, evaluating and debugging AI applications and agents in your own environment.

github

Capabilities: Trace agent runs and tool calls using OpenTelemetry and OpenInference standards · Run LLM-as-a-judge and custom evaluations on traces · Compare prompts and datasets in experiments and manage prompts · Self-host in a notebook, Docker or Kubernetes

Interfaces: Web UI · Python SDK · REST API · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Zfinix

⭐ 107 stars · Growth: 9d +15 stars (+16.3%)

🤖 AI Agents

Aster is a local-first software-engineering agent that uses shared tools, permissions, memory, skills, and retrieval for chat, code review, and fixes.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @AtomicBot-ai

⭐ 2,740 stars · Growth: 9d +238 stars (+9.5%)

🤖 AI Agents | 💻 Local Inference | 💻 Terminal-Friendly

Atomic Agent is a local-first terminal coding assistant that runs open-weight models through llama.cpp and keeps its control loop and state on the user’s machine. It can browse, work with files and Git, retain memory across sessions, and connect external tools over MCP.

github

Capabilities: Local model inference · Browser, file, and Git tools · Persistent memory and MCP tool connections

Interfaces: CLI · TUI · HTTP API · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @agentrhq

⭐ 93 stars · Growth: 9d +1 stars (+1.1%)

🤖 AI Agents | 🛡️ Safety Guardrails (Safeguarding)

A local credential broker for AI agents that stores credentials in an encrypted vault and injects them through a local HTTPS proxy, keeping raw secrets away from agent processes.

github

Capabilities: Manage OAuth2 and API key credentials for agent tools · Store credentials in an encrypted local vault · Inject credentials into requests through a local HTTPS proxy

Interfaces: Python package · Local HTTPS proxy · CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @phodal

⭐ 4,552 stars · Growth: 9d +6 stars (+0.1%)

🤖 AI Agents

AutoDev is an AI-powered coding assistant offering multilingual support, automatic code generation, and debugging assistance, featuring customizable prompts and specialized tools for development, testing, documentation, and the integration of custom AI agents, with a focus on experimenting and building AI agents using its UI framework

github | docs

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @microsoft

⭐ 61,251 stars · Growth: 9d +87 stars (+0.1%)

⚙️ Development Frameworks

AutoGen is a multi-agent conversation framework facilitating the development of next-gen LLM applications, highlighted by various accomplishments and offering enhanced LLM inferences, customizable agents, and comprehensive documentation

github | video

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @microsoft

⭐ 61,251 stars · Growth: 9d +87 stars (+0.1%)

⚙️ Development Frameworks

AutoGen Studio 2.0 is Microsoft's advanced AI development tool, offering a user-friendly interface, powerful Python API, and comprehensive features for creating and controlling AI agents and workflows

github | website | landing page | research paper

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Significant-Gravitas

⭐ 187,644 stars · Growth: 9d +87 stars (+0.0%)

🤖 AI Agents

AutoGPT provides accessible AI tools for building and using AI agents, offering a comprehensive framework including Forge for agent creation, agbenchmark for performance evaluation, a leaderboard for competition, a user-friendly UI, and CLI for seamless integration and management

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @autohandai

⭐ 202 stars · Growth: 9d +3 stars (+1.5%)

⚙️ Development Frameworks

Autohand Code CLI is a terminal-native AI coding agent that plans and edits code, runs tools, and supports testing, memory, skills, and permission controls.

github

Capabilities: code planning and editing · tool execution and testing · project memory and skills

Interfaces: CLI

Submitted by: @x-zheng16 submission accepted PR · Maintained by: not yet verified · Repository owner: @x-zheng16

⭐ 151 stars · Growth: 9d +8 stars (+5.6%)

🤖 AI Agents

An open survey of embodied AI safety research, organizing papers on perception, cognition, planning, interaction, and agentic-system risks.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @mergisi

⭐ 3,995 stars · Growth: 9d +10 stars (+0.3%)

🤖 AI Agents

A community-maintained collection of OpenClaw agent templates with personas, tool bindings, memory rules, and operating schedules.

github

Capabilities: Browse reusable OpenClaw agent configurations across roles · Adapt templates with persona, tool, and memory instructions · Start new role-specific agents from Markdown configurations

Interfaces: Markdown templates · OpenClaw

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Necmttn

⭐ 115 stars · Growth: 9d +3 stars (+2.7%)

👁️ Observability Frameworks | 💻 Terminal-Friendly

A local agent-experience graph for coding agents that captures sessions, recalls project context, and turns recurring workflow friction into reviewable improvements.

github

Capabilities: Capture coding-agent session telemetry · Recall project-specific session context · Track costs and workflow outcomes · Propose reviewable workflow improvements

Interfaces: CLI · Local Studio UI · Claude Code integration · Codex integration

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @iPythoning

⭐ 189 stars · Growth: 9d +1 stars (+0.5%)

🤖 AI Agents

An OpenClaw-based AI sales development representative template for managing B2B export leads across a multi-stage sales pipeline.

github

Capabilities: Capture and score B2B prospects · Manage a staged sales pipeline and follow-ups · Coordinate prospect communication across messaging and email channels

Interfaces: OpenClaw agent template · WhatsApp · Telegram · Email

Submitted by: @ct-jaryn submission accepted PR · Maintained by: not yet verified · Repository owner: @chaitin

⭐ 101 stars · Growth: 9d +1 stars (+1.0%)

🔧 Tool Calling (Function Calling)

The official open-source client integration for Baizhi Cloud's hosted MCP service, with a Gemini CLI extension, registry manifest, and setup guides for web search, webpage retrieval, and structured extraction. The backend remains hosted and requires an API key.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @BoundaryML

⭐ 9,381 stars · Growth: 9d +97 stars (+1.0%)

🏗️ Structured Outputs

BAML is a domain-specific language for building typed LLM functions with structured outputs and streaming.

github

Submitted by: @maxfain (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @maxfain

⭐ 2 stars · Growth: 9d +2 stars (+0.0%)

🤖 AI Agents

Open task marketplace for AI agents to discover, claim, and deliver work using signed receipts and optional USDC bounties.

github

Capabilities: Agent identity and reputation · Task discovery and exchange · Signed task receipts · USDC task payments

Interfaces: Web UI · HTTP API · TypeScript SDK and CLI · Python SDK · MCP server

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sagentic-ai

⭐ 79 stars · Growth: 9d +1 stars (+1.3%)

⚙️ Development Frameworks

Bazed Agent Framework, aimed at empowering developers to build autonomous agent swarms without requiring deep Python ML knowledge, is facilitating the creation of sophisticated systems through TypeScript for enhanced autonomy and reliability

github | website | discord

Submitted by: @allxsmith (self-reported founder/team) submission accepted PR · Maintained by: @allxsmith (self-reported) evidence · Repository owner: @allxsmith

⭐ 12 stars

🔌 MCP Servers | 🖥️ UI Development

bestax-mcp is an offline MCP server that provides coding agents with Bestax React component APIs, examples, CSS variables, helper props, and task-specific skills.

github

Capabilities: Look up component props, examples, and CSS variables · Serve seven UI-development skills and their references offline

Interfaces: MCP server

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @BindBrain

⭐ 6 stars · Growth: 9d +2 stars (+50.0%)

⚙️ Development Frameworks

BindAI is a modular Python framework for building AI applications with reusable components for agents, tools, workflows, memory, knowledge retrieval, model providers, and integrations.

github | website | docs

Submitted by: @venkat22022202 submission accepted PR · Maintained by: not yet verified · Repository owner: @venkat22022202

⭐ 9 stars · Growth: 9d +1 stars (+12.5%)

🤖 AI Agents

Self-hostable API key proxy for AI agents with per-token access controls, budget limits, and an OpenAI-compatible gateway for multiple model providers.

github | demo

Submitted by: @Blave-Wei (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @Blave-TW

⭐ 88 stars · Growth: 9d +55 stars (+166.7%)

🤖 AI Agents

Blave Agent is a macOS quant research workspace that connects coding agents such as Claude Code or Codex to help develop and backtest trading strategies, including Monte Carlo permutation tests, parameter scans, and walk-forward analysis. Live trading runs through a separate deterministic process with position reconciliation and a kill switch.

github

Capabilities: Agent-assisted strategy development · Backtesting and robustness analysis · Deterministic trading automation

Interfaces: macOS desktop app · Claude Code and Codex integration

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @browser-use

⭐ 117,092 stars · Growth: 9d +811 stars (+0.7%)

🌐 Web Browsing Frameworks

browser-use lets AI agents interact with websites through a browser using page structure and vision.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @PramodDutta

⭐ 3 stars · Growth: 9d +1 stars (+50.0%)

📊 Evaluation Frameworks

BrowserBash is a CLI and MCP server for running plain-English browser tests in Chrome and returning deterministic assertions and exit codes for agent-driven UI validation.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @VRSEN

⭐ 4,590 stars · Growth: 9d +10 stars (+0.2%)

🤖 AI Agents

BrowsingAgent, an AI web navigation tool, has been integrated into the Agency Swarm framework to enable human-like browsing capabilities for automated AI operations

github | video

Submitted by: @azeemkafridi submission accepted PR · Maintained by: not yet verified · Repository owner: @azeemkafridi

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🔧 Tool Calling (Function Calling)

Open-source Python and Node.js SDKs and an MCP server that connect agents to BulkPublish's hosted social publishing API, with tools for drafting, scheduling, publishing, approvals, and analytics.

github

Submitted by: @ranglang submission accepted PR · Maintained by: not yet verified · Repository owner: @busabase

⭐ 321 stars

🔌 MCP Servers

Busabase is an open-source local first database and workspace that exposes structured records, documents, skills and workspace operations to external AI agents through MCP, OpenAPI and a CLI, with every write carried as a reviewable change request.

github

Capabilities: Expose structured records, documents, reusable skills and workspace operations to agents · Let agents propose data or document changes as change requests with a diff, author, message and history · Merge changes immediately or hold them for review depending on the credential permission

Interfaces: MCP · OpenAPI · CLI

Submitted by: @BuyWhere submission accepted PR · Maintained by: not yet verified · Repository owner: @BuyWhere

⭐ 15 stars · Growth: 9d +1 stars (+7.1%)

🔧 Tool Calling (Function Calling)

An open-source MCP client that connects AI agents to BuyWhere’s hosted product-data API for product search, price comparison, and deal discovery.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @calcom

⭐ 48,851 stars · Growth: 9d +203 stars (+0.4%)

🤖 AI Agents

Cal.ai is an open-source AI scheduling assistant that manages email communications for booking, rearranging, and inquiring about meetings, leveraging a LangChain Agent Executor and MailParser for efficient scheduling without API key exposure

website | github

Submitted by: @skeptrunedev (self-reported founder/team) submission accepted PR · Maintained by: @skeptrunedev (self-reported) evidence · Repository owner: @skeptrunedev

⭐ 3 stars

📞 Phone Calling

call4me is a voice agent that places user-requested phone calls to businesses through an MCP server, navigates phone menus, and returns the call outcome and transcript.

github

Capabilities: Place and schedule calls to businesses · Navigate phone menus and return transcripts and outcomes

Interfaces: MCP server · Coding-agent integration

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @camel-ai

⭐ 17,809 stars · Growth: 9d +35 stars (+0.2%)

⚙️ Development Frameworks

CAMEL (Communicative Agents for Mind Exploration of Large Language Model Society) is an open-source library designed for studying autonomous and communicative agents, facilitating research in understanding their behaviors, capabilities, and potential risks through scalable techniques and cooperative frameworks, including role-playing, with extensive documentation, examples, and datasets, while also supporting integration with open-source models as backends for diverse applications

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Ceki-me

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🔌 MCP Servers

Ceki's open-source MCP server connects agents to a hosted marketplace for real-browser sessions and human services, with companion Python and JavaScript SDKs.

github | github | github

Submitted by: @islomkhon (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @islomkhon

⭐ 4 stars

🤖 AI Agents

CerebrumKit is a self-hosted starting point for agentic projects where agents are assembled in an admin panel from skills and tools stored as database rows, with a workflow canvas and a client chat panel.

github

Capabilities: Assemble agents from skills and Python tools with OpenAI function specs, editable without a restart · Route work through a start, group and stop workflow canvas, with a second group reviewing the first and one agent delegating to another · Point agents at any OpenAI compatible endpoint including DeepSeek, OpenAI, OpenRouter, Groq, vLLM and Ollama

Interfaces: Web UI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @OpenBMB

⭐ 34,439 stars · Growth: 9d +43 stars (+0.1%)

⚙️ Development Frameworks

ChatDev is a virtual software company utilizing intelligent agents to revolutionize the digital world through programming, offering a highly customizable framework and integrating innovative approaches like Experiential Co-Learning, Docker support, Git management, and Human-Agent Interaction

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @anthropics

⭐ 8,213 stars · Growth: 9d +50 stars (+0.6%)

⚙️ Development Frameworks

Anthropic's Python SDK for building agents with Claude using the Claude Code harness.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @TIGER-AI-Lab

⭐ 932 stars · Growth: 9d +83 stars (+9.8%)

📊 Evaluation Frameworks

ClawBench is an open-source benchmark for browser agents, with everyday tasks across live websites, execution traces, and request-level outcome evaluation.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @vivekchand

⭐ 424 stars · Growth: 9d +4 stars (+1.0%)

👁️ Observability Frameworks

ClawMetry is an open-source, local-first dashboard that reads existing session files from OpenClaw, NVIDIA NemoClaw, Goose, and Qwen Code. It provides agent session replay, tool-call timelines, and token and cost reporting where runtimes expose that data.

github

Capabilities: Agent session monitoring and replay · Tool-call timelines · Token and cost reporting

Interfaces: CLI · Local web dashboard

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @cline

⭐ 69,807 stars · Growth: 9d +495 stars (+0.7%)

🤖 AI Agents

Cline is an open-source coding agent available in IDEs, terminals, and desktop, where it can plan and carry out software development tasks with user control.

github

Capabilities: code planning and editing · tool use across development workflows · MCP integrations

Interfaces: IDE extension · CLI · Desktop app

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @CogitoForge-AI

⭐ 21 stars · Growth: 9d +2 stars (+10.5%)

👤 Personal Assistants | 🖥️ UI Development

Cogito Studio is a cross-platform desktop AI workspace built with Tauri, React, and Rust, offering multi-provider model access, MCP integration, workspace management, and local SQLite storage.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @topoteretes

⭐ 31,338 stars · Growth: 9d +353 stars (+1.1%)

⚙️ Development Frameworks

Cognee is an open-source framework aimed at simplifying data processing for large language models (LLMs) by creating knowledge graphs and data models, offering tools for information addition, knowledge creation, and similarity-based search

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @openonion

⭐ 1,493 stars · Growth: 9d +14 stars (+0.9%)

⚙️ Development Frameworks | 🔧 Tool Calling (Function Calling)

A Python CLI harness that lets coding agents access local tools such as Gmail, a browser, and files from their shell.

github

Capabilities: Give coding agents shell-accessible tools for email and browser workflows · Keep user credentials on the local machine · Build agent workflows by composing command-line capabilities

Interfaces: Python package · CLI · Local shell tools

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @upstash

⭐ 62,651 stars · Growth: 9d +225 stars (+0.4%)

🔌 MCP Servers

Context7 provides version-specific library documentation and code examples to agents through an MCP server.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @contextstream

⭐ 44 stars · Growth: 9d +1 stars (+2.3%)

🧠 Long-Term Memory

ContextStream is an MCP server for shared project context and persistent agent memory, with semantic search, traceable sources, and scoped access.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @continuedev

⭐ 36,105 stars · Growth: 9d +77 stars (+0.2%)

🤖 AI Agents

Continue is an open-source autopilot plugin for VS Code and JetBrains, enhancing coding with LLMs through features like task and tab autocomplete, natural language edits, file generation, and customization options, available under the Apache 2.0 license

github | website

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @CorellisOrg

⭐ 31 stars · Growth: 9d +1 stars (+3.3%)

🤖 AI Agents | 🧠 Long-Term Memory | 🛡️ Safety Guardrails (Safeguarding)

An OpenClaw-based multi-agent workforce framework with shared team memory, goal decomposition, fleet learning, task management, and approval workflows.

github

Capabilities: Decompose team goals and coordinate tasks across agent fleets · Share searchable knowledge and promote lessons across agents · Route sensitive actions through human approval workflows

Interfaces: OpenClaw · Docker · Slack · Agent skills

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @agentweave

⭐ 13 stars · Growth: 9d +1 stars (+8.3%)

⚙️ Development Frameworks

An open protocol for coordinating AI agent teams through shared Markdown task files and a chief-of-staff workflow.

github

Capabilities: Coordinate chief-of-staff and worker agents through shared task files · Track team tasks and agent status across projects · Integrate with multiple coding-agent runtimes

Interfaces: Markdown protocol · Claude Code plugin · Runtime adapters

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @SKULLFIRE07

⭐ 9 stars · Growth: 9d +1 stars (+12.5%)

🧠 Long-Term Memory

A local-first memory tool for AI coding assistants that captures project context and carries it across sessions.

github

Capabilities: Capture project decisions, patterns, and context across coding sessions · Persist and retrieve multi-layer project memory · Inject relevant memory into new assistant sessions

Interfaces: VS Code extension · CLI · MCP server

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @CorvinLabs

⭐ 14 stars · Growth: 9d +3 stars (+27.3%)

🤖 AI Agents | 💻 Local Inference | 💻 Operating System (OS)

CorvinOS is a self-hosted agentic OS that connects local and cloud models to messaging platforms including Discord, Telegram, WhatsApp, Slack, and email.

github

Submitted by: @foklepoint (self-reported founder/team) submission accepted PR · Maintained by: @foklepoint (self-reported) evidence · Repository owner: @foklepoint

⭐ 2 stars

🔌 MCP Servers

A hosted MCP server that answers U.S. federal and local court filing-rule questions from cited source documents, checks judge-specific filing requirements, and looks up court holidays and deadlines.

github

Capabilities: Search filing rules and judge standing orders with source citations · Check briefs against judge-specific filing requirements · Look up court holidays and court deadlines

Interfaces: MCP · Streamable HTTP

Submitted by: @BunsDev (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @OpenCoven

⭐ 66 stars

💻 Terminal-Friendly

Coven is an open-source Rust daemon and CLI that runs Codex, Claude Code and other coding agent harnesses as PTY sessions inside explicit project root boundaries, with session history kept in SQLite.

github

Capabilities: Run coding agent sessions inside explicit project root boundaries and reject working directories that escape them · Persist session history and events in SQLite after a session ends · Attach to and coordinate sessions from the CLI or other clients over a versioned local socket API

Interfaces: CLI · Local socket API

Submitted by: @benjaminLedel (self-reported founder/team) submission accepted PR · Maintained by: @benjaminLedel (self-reported) evidence · Repository owner: @benjaminLedel

⭐ 4 stars

🤖 AI Agents | 🔄 Flow Engineering (Platform Engineering)

A self-hosted platform for organizational AI agents with isolated sandboxes, per-agent credentials, backlogs and wake sources, human approvals, run logs, and cost tracking.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ZiwayZhao

⭐ 21 stars · Growth: 9d +1 stars (+5.0%)

⚙️ Development Frameworks | 📏 Standardization

A peer-to-peer protocol for serving and invoking agent skills over XMTP with encrypted communication and revocable trust.

github

Capabilities: Serve remote agent skills without distributing their implementation · Delegate tasks asynchronously between agents · Apply trust levels and version checks to remote skill calls

Interfaces: Python package · MCP server · XMTP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @unclecode

⭐ 84,721 stars · Growth: 9d +456 stars (+0.5%)

🌐 Web Browsing Frameworks

Crawl4AI is an open-source web crawler that produces Markdown and structured data for AI applications.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @crewAIInc

⭐ 59,330 stars · Growth: 9d +310 stars (+0.5%)

⚙️ Development Frameworks

crewAI is a cutting-edge AI framework designed for orchestrating role-playing, autonomous AI agents, enabling seamless collaboration and complex task handling

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @momenbasel

⭐ 9 stars · Growth: 9d +1 stars (+12.5%)

🤖 AI Agents | 🛡️ Safety Guardrails (Safeguarding)

A pre-transaction safety hook for AI agents that assesses crypto contract and token risks using security-oracle data and bytecode analysis.

github

Capabilities: Check smart contracts for honeypot, blacklist, and rug-pull indicators · Aggregate risk signals from multiple security data sources · Add a review gate before agent-initiated crypto transactions

Interfaces: Python package · Claude Code hook · Agent skill

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @cuga-project

⭐ 888 stars · Growth: 9d +8 stars (+0.9%)

⚙️ Development Frameworks

CUGA is an enterprise agent harness for configuring and deploying generalist agents across API and web workflows, with tool orchestration, safety policies, and evaluation.

github

Capabilities: enterprise workflow automation · tool orchestration across APIs and web apps · policy and evaluation controls

Interfaces: Python SDK · API · Web UI

Submitted by: not yet recovered from repository history · Maintained by: @amavashev evidence · Repository owner: @runcycles

⭐ 5 stars · Growth: 9d +1 stars (+25.0%)

🤖 AI Agents | 🛡️ Safety Guardrails (Safeguarding) | 📏 Standardization

A self-hosted runtime authority for AI agents that enforces budgets, risk limits, and tool-action policies before execution through a concurrency-safe reservation lifecycle.

github

Capabilities: Enforce agent spend budgets and action policies before execution · Reserve, commit, or release usage with concurrency-safe controls · Provide tenant-isolated governance through the Cycles Protocol

Interfaces: Java server · REST API · MCP server · Docker

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @eosphoros-ai

⭐ 20,077 stars · Growth: 9d +31 stars (+0.2%)

⚙️ Development Frameworks

DB-GPT revolutionizes database interactions using private LLM technology, enabling streamlined AI-native data app development with multi-model management, Text2SQL optimization, and fine-tuning, facilitating enterprises and developers to create bespoke applications in the Data 3.0 era

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @VRSEN

⭐ 4,590 stars · Growth: 9d +10 stars (+0.2%)

🤖 AI Agents

Devid Agent, a new AI software development tool, has been integrated into the Agency Swarm framework to enhance automated AI agency operations, alternative to Cognition AI's Devin

github | video

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @kuafuai

⭐ 5,967 stars · Growth: 9d +2 stars (+0.0%)

⚙️ Development Frameworks

DevOpsGPT is an AI-driven software development automation solution that combines large language models with DevOps tools to convert natural language requirements into working software, enhancing development efficiency, shortening cycles, and reducing communication costs

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Charlie85270

⭐ 350 stars · Growth: 9d +3 stars (+0.9%)

🤖 AI Agents | ⚙️ Development Frameworks

A desktop workspace for orchestrating, monitoring, and automating multiple AI coding agents.

github

Capabilities: Run and monitor multiple coding-agent CLIs from one desktop app · Manage agent work with Kanban boards and scheduled automations · Control agent sessions remotely through connected messaging tools

Interfaces: Desktop app · MCP servers

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Moresyl

⭐ 76 stars · Growth: 9d +49 stars (+181.5%)

🖥️ UI Development

DSH Studio is a cross-platform desktop interface for supervising and managing the local DeepSeek Harness service. It monitors and restarts the service, manages plugin installation, and offers optional paired-phone access through a separate local-network gateway.

github

Capabilities: Local agent-runtime supervision · Plugin installation and management · Paired-phone access to the desktop session

Interfaces: Desktop app · Paired mobile browser

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @stanfordnlp

⭐ 38,491 stars · Growth: 9d +207 stars (+0.5%)

🔄 Flow Engineering (Platform Engineering)

A cutting-edge framework that compiles declarative language model calls into self-improving pipelines, enabling the systematic and efficient optimization of LM prompts and weights within complex systems

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @zoharbabin

⭐ 115 stars · Growth: 9d +4 stars (+3.6%)

🤖 AI Agents

A local forensic M&A due-diligence application that coordinates specialist agents across nine domains and traces findings to source pages and quotations.

github

Capabilities: Analyze data-room documents across specialist domains · Cross-reference findings across legal, finance, commercial, and technical topics · Generate cited HTML and Excel reports

Interfaces: CLI · Python package · Docker

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @e2b-dev

⭐ 14,143 stars · Growth: 9d +173 stars (+1.2%)

⚙️ Development Frameworks

E2B Sandbox offers secure cloud environments tailored for AI agents and apps, facilitating long-running sessions with various tools and can be integrated with any large language model

github | github profile

Submitted by: @paperandbeyond23-gif submission accepted PR · Maintained by: not yet verified · Repository owner: @paperandbeyond23-gif

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🔧 Tool Calling (Function Calling)

Open-source agent skills and a lightweight MCP stdio bridge for Edgrapi’s hosted API, which returns normalized SEC and US government data as JSON.

github

Submitted by: @vakandi (self-reported founder/team) submission accepted PR · Maintained by: @vakandi (self-reported) evidence · Repository owner: @vakandi

⭐ 2 stars

🤖 AI Agents | 🧠 Long-Term Memory

An OpenCode-based runtime for scheduled, persistent subagents, each with an isolated workspace, plus fleet status, streaming logs, retries, and restart handling.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Mafengwo292

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🤖 AI Agents

EVO-AI is a distributed language-model project that evolves model weights across volunteer nodes and exposes generation through HTTP and A2A JSON-RPC interfaces.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @mediar-ai

⭐ 341 stars · Growth: 9d +2 stars (+0.6%)

🤖 AI Agents | 💻 Operating System (OS)

A local macOS AI computer agent that uses voice and visual context to operate apps, browsers, documents, and workflows.

github

Capabilities: Control macOS apps and browser workflows using natural language · Understand on-screen context while carrying out desktop tasks · Create and edit code and documents through computer use

Interfaces: macOS desktop app · Voice interface · ACP bridge

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @firecrawl

⭐ 188,355 stars · Growth: 9d +3,672 stars (+2.0%)

🌐 Web Browsing Frameworks

Firecrawl turns websites into Markdown and structured data through crawl, scrape, and extract APIs.

github

Submitted by: not yet recovered from repository history · Maintained by: @moraispgsi evidence · Repository owner: @synergenius-fw

⭐ 2 stars · Growth: 9d +2 stars (+0.0%)

🤖 AI Agents | ⚙️ Development Frameworks

A deterministic TypeScript workflow compiler that turns JSDoc-annotated functions into standalone workflow code you can inspect, test, and own.

github

Capabilities: Validate and compile annotated TypeScript workflows into execution graphs · Run workflows locally and inspect execution steps in a web console · Create and operate workflows through MCP tools and approval gates

Interfaces: TypeScript package on npm · CLI · MCP server · Local web console

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @FlowiseAI

⭐ 55,488 stars

⚙️ Development Frameworks

Flowise simplifies the creation of applications leveraging large language models (LLMs) by providing a drag-and-drop interface for customizing AI workflows, offering easy installation, Docker support, development tools, and documentation for integrating various functionalities such as authentication, streaming, and custom tools to enhance AI agents' capabilities

github | website | docs | github profile

Submitted by: @BlueHotDog submission accepted PR · Maintained by: not yet verified · Repository owner: @frontman-ai

⭐ 709 stars · Growth: 9d +2 stars (+0.3%)

🤖 AI Agents | 🖥️ UI Development

A browser-based AI coding agent that uses live application context to edit frontend source files with hot reload through Next.js, Astro, and Vite integrations.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @blackwell-systems

⭐ 56 stars · Growth: 9d +5 stars (+9.8%)

📏 Standardization | 🔧 Tool Calling (Function Calling)

An AI-native structured-data wire format with implementations in six languages and a proxy for adopting compact agent tool responses.

github

Capabilities: Encode and decode structured data · Reduce token use in agent tool responses · Adopt compact responses through a proxy

Interfaces: Go SDK · Rust SDK · TypeScript SDK · Python SDK · Swift SDK · Kotlin SDK

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @giselles-ai

⭐ 556 stars · Growth: 9d +2 stars (+0.4%)

🤖 AI Agents

Giselle is an open-source visual studio for building and running AI agent workflows that combine models, tools, and data sources to deliver products.

github

Capabilities: visual agent workflow design · multi-model and data source integration · workflow deployment

Interfaces: Web UI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @github

⭐ 33,352 stars · Growth: 9d +155 stars (+0.5%)

🔌 MCP Servers | 🔧 Tool Calling (Function Calling)

GitHub's official MCP server lets AI agents work with repositories, issues, pull requests, Actions, and code security through the Model Context Protocol.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @alexk-dev

⭐ 41 stars · Growth: 9d +1 stars (+2.5%)

⚙️ Development Frameworks

GolemCore Bot is an agent platform for AI-native teams, with channel-connected agents, skills, plugins, MCP servers, memory, and execution tracing.

github

Capabilities: channel-connected agent runtime · skills, plugins, and MCP integration · persistent memory and execution tracing

Interfaces: Web UI · Telegram · Webhooks

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @google

⭐ 21,701 stars · Growth: 9d +56 stars (+0.3%)

⚙️ Development Frameworks

Google's open-source Agent Development Kit for building, evaluating, and deploying multi-agent systems.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Pythagora-io

⭐ 33,659 stars · Growth: 9d -3 stars (-0.0%)

🤖 AI Agents

GPT Pilot is an open-source AI developer tool that aims to provide a comprehensive development companion, capable of writing features, debugging, and interacting with users, presenting itself as an alternative to Devin, the world's first AI software engineer developed by Cognition Labs

github | discord

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @assafelovic

⭐ 29,902 stars · Growth: 9d +279 stars (+0.9%)

🤖 AI Agents

GPT Researcher is an AI-powered autonomous agent designed for efficient and unbiased online research, generating detailed reports by leveraging recent advancements in AI and web scraping, with a focus on speed, reliability, and cost-effectiveness

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @gptme

⭐ 4,443 stars · Growth: 9d +14 stars (+0.3%)

🤖 AI Agents | ⚙️ Development Frameworks | 💻 Terminal-Friendly

A provider-agnostic personal AI agent that runs in terminals from laptops and remote servers, with local tools for coding and general knowledge work.

github

Capabilities: Use shell, Python, web, and vision tools to complete tasks · Run persistent autonomous agents across local and headless environments · Extend agents with plugins, skills, MCP, and ACP integrations

Interfaces: Terminal CLI · Python package · MCP · ACP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @guardrails-ai

⭐ 7,482 stars · Growth: 9d +32 stars (+0.4%)

🛡️ Safety Guardrails (Safeguarding)

Guardrails is a Python framework for building reliable AI applications, offering Input/Output Guards to detect and mitigate risks, along with structured data generation from large language models (LLMs)

github | twitter

Submitted by: not yet recovered from repository history · Maintained by: @Harsha-Nori evidence, @marcotcr evidence, @slundberg evidence · Repository owner: @guidance-ai

⭐ 21,788 stars · Growth: 9d +11 stars (+0.1%)

🛡️ Safety Guardrails (Safeguarding)

The text describes 'guidance,' a programming paradigm that enhances control and efficiency in model generation by allowing for constraints like regex and CFGs, integrating stateful control, and offering a simplified interface for complex generation scenarios

github | docs

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @HagiCode-org

⭐ 5 stars · Growth: 9d +1 stars (+25.0%)

⚙️ Development Frameworks | 🖥️ UI Development

An Electron desktop app for installing, running, monitoring, and updating HagiCode Server, with local resource monitoring and agent executor selection.

github

Capabilities: Set up and manage HagiCode Server · Monitor local resources and service health · Select coding-agent executors

Interfaces: Desktop UI · Claude Code · Codex · GitHub Copilot CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @bkuan001

⭐ 82 stars · Growth: 9d +2 stars (+2.5%)

👁️ Observability Frameworks

halo-record creates tamper-evident, hash-chained records of AI agent tool calls, model calls, and data access, with verifiable reports and policy checks.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @deepset-ai

⭐ 26,648 stars · Growth: 9d +49 stars (+0.2%)

⚙️ Development Frameworks

Haystack is an end-to-end LLM framework facilitating the construction of applications powered by LLMs, Transformer models, vector search, and more, offering flexibility, transparency, and extensibility, with features including retrieval-augmented generation, document search, question answering, and semantic search, along with a diverse user base including companies like Airbus, Apple, and Netflix

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Helicone

⭐ 6,199 stars · Growth: 9d +24 stars (+0.4%)

👁️ Observability Frameworks

Helicone is an open-source observability platform for Language Learning Models (LLMs), providing features like request logging, caching, rate limiting, cost and latency tracking, UI-based prompt iteration, and collaboration tools

website | github

Submitted by: @slavakurilyak submission accepted PR · Maintained by: not yet verified · Repository owner: @NousResearch

⭐ 251,029 stars

🤖 AI Agents | 🧠 Long-Term Memory | 💻 Terminal-Friendly

Hermes Agent is an open-source agent from Nous Research that runs in a terminal and through messaging platforms, keeps memory and creates skills from experience, and works with the model provider of your choice.

github

Capabilities: Run tasks from a terminal interface with slash commands, conversation history and streaming tool output · Serve one agent through Telegram, Discord, Slack, WhatsApp, Signal and the CLI from a single gateway process · Keep agent curated memory, create skills after complex tasks, and search past sessions · Run scheduled automations and spawn isolated subagents for parallel work · Execute on local, Docker, SSH, Singularity, Modal, Daytona or Vercel Sandbox backends

Interfaces: Terminal UI · CLI · Messaging gateway · Desktop app

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @fresh-claw

⭐ 18 stars · Growth: 9d +2 stars (+12.5%)

👤 Personal Assistants

A Chinese language and workflow enhancement pack for Hermes Agent desktop, TUI, and CLI, with localized prompts, gateway guidance, pairing approvals, and installation tools.

github

Capabilities: Localize Hermes Agent interfaces and prompts into Chinese · Provide setup guidance for messaging gateways and pairing approvals · Install version-matched Chinese enhancement packages

Interfaces: Hermes Agent Desktop · TUI · CLI · Windows PowerShell · Shell installer

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @JesseGdotIO

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🔌 MCP Servers | 🛡️ Safety Guardrails (Safeguarding)

The open-source Hermes Plant MCP server connects agents to action safety and commerce assurance services for shell, Git, SQL, deployment, and x402 workflows.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @aden-hive

⭐ 11,091 stars · Growth: 9d +19 stars (+0.2%)

⚙️ Development Frameworks

A model-agnostic runtime for production AI agent teams, with dynamic worker colonies, shared task state, and human oversight.

github

Capabilities: Grow specialized agent teams to carry out long-running processes · Coordinate workers through a shared ledger and persistent task plans · Recover from failures and manage cost, observability, and human review

Interfaces: Python runtime · CLI · MCP tools

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @MarlBurroW

⭐ 68 stars · Growth: 9d +6 stars (+9.7%)

🤖 AI Agents | 👤 Personal Assistants

Hivekeep is a self-hosted platform of persistent AI agents that remember, collaborate, and respond across connected channels, packaged to run on your own server.

github

Capabilities: persistent agent memory · multi-agent collaboration · messaging channel integrations

Interfaces: Web UI · API

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @hivemoot

⭐ 17 stars · Growth: 9d +1 stars (+6.2%)

🤖 AI Agents

Hivemoot runs a team of AI agents on GitHub repositories, where agents propose work, debate in issues, implement changes, review pull requests, and vote on decisions.

github

Capabilities: autonomous software development team · issue and pull request collaboration · agent governance and voting

Interfaces: GitHub App · CLI · Web UI

Submitted by: not yet recovered from repository history · Maintained by: @HGraphPunks evidence, @kantorcodes evidence, @rocketmay evidence · Repository owner: @hashgraph-online

⭐ 1,240 stars · Growth: 9d +11 stars (+0.9%)

⚙️ Development Frameworks | 📏 Standardization

A TypeScript SDK implementing Hashgraph Online Consensus Standards, with Registry Broker clients for agent discovery and chat.

github

Capabilities: Implement and consume Hashgraph Online agent standards · Search for agents through the Registry Broker · Integrate agent chat and registry workflows

Interfaces: TypeScript package · Registry Broker API

Submitted by: @rushar-labs (self-reported founder/team) submission accepted PR · Maintained by: @rushar-labs (self-reported) evidence · Repository owner: @rusharlabs

⭐ 5 stars

📊 Evaluation Frameworks

House Party Protocol is a local first Python harness for Claude Code and Codex CLI teams that records criterion command results outside the model, refuses verdicts from the builder's own lane, and binds approvals to a spec hash.

github

Capabilities: Run a criterion command with its exit code recorded outside the model and its artifacts hashed · Refuse an attestation from the builder own lane or model family and bind approvals to the spec hash, base commit and file snapshot · Report pass@k and pass^k separately over repeated runs and gate on the result

Interfaces: CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @human-pages-ai

⭐ 8 stars · Growth: 9d +1 stars (+14.3%)

🔌 MCP Servers

Human Pages MCP Server lets AI agents find, hire, and coordinate people for tasks that require real-world human work, including testing, localization, and community operations.

github

Capabilities: search people by skill and location · create jobs and manage messaging · human task fulfillment

Interfaces: MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @YugantM

⭐ 6 stars · Growth: 9d +1 stars (+20.0%)

🛡️ Safety Guardrails (Safeguarding)

An open-source trust registry that aggregates public evidence about AI agents and MCP servers, with project comparisons, score history, and machine-readable data.

github

Capabilities: Browse evidence about AI-agent and MCP-server projects · Compare project trust signals · Review score history and evidence coverage · Consume machine-readable registry data

Interfaces: Web app · Public JSON API · Embeddable badges

Submitted by: @louis030195 (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @louis030195

⭐ 5 stars

🧠 Long-Term Memory | 🔌 MCP Servers

Hyperconsciousness is an MIT-licensed developer alpha Rust knowledge store for humans and AI agents that keeps signed, encrypted, append-only records and exposes scoped, expiring retrieval and capture grants through CLI, MCP and HTTP.

github

Capabilities: Keep signed, encrypted, append-only records and sync them across devices · Let agents retrieve authorized notes across sessions or append new knowledge through MCP using bounded, scoped and expiring grants · Install ingestion skills that capture chats and business tool data as records

Interfaces: CLI · MCP · HTTP API

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @567-labs

⭐ 13,972 stars · Growth: 9d +30 stars (+0.2%)

🏗️ Structured Outputs

Instructor, a Python library, facilitates working with structured outputs from large language models (LLMs), offering features like response model specification, retry management, validation, and streaming support, primarily aimed at enhancing workflows of AI agents utilizing LLMs

github | website

Submitted by: @douglasborthwick-crypto submission accepted PR · Maintained by: not yet verified · Repository owner: @insumerapi

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🤖 AI Agents

Open-source MCP and LangChain integrations for calling Insumer’s hosted service to verify on-chain wallet conditions and return signed attestations.

github | github

Capabilities: On-chain wallet condition verification · Cryptographically signed attestations · Privacy-preserving balance checks

Interfaces: MCP server · LangChain toolkit

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @janhq

⭐ 44,781 stars · Growth: 9d +130 stars (+0.3%)

💻 Local Inference

Jan is an open-source, development-stage ChatGPT alternative that operates fully offline on diverse hardware platforms, supporting universal architectures from PCs to multi-GPU clusters

github | github profile

Submitted by: @emarco177 (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @everafterlabs

⭐ 10 stars

🛡️ Safety Guardrails (Safeguarding)

jes is an open-source guardrails library for checking agent inputs, retrieved content, tool calls, tool results, and responses for prompt injection, jailbreaks, and sensitive-data leaks.

github

Capabilities: Check content at agent trust boundaries · Redact configured secrets and personal data locally · Block risky prompts and tool calls through coding-agent hooks

Interfaces: Python library · Coding-agent hooks

Submitted by: @IRONICBo (self-reported founder/team) submission accepted PR · Maintained by: @IRONICBo (self-reported) evidence · Repository owner: @socai-io

⭐ 144 stars · Growth: 9d +62 stars (+75.6%)

🤖 AI Agents

A browser-grounded social research agent that lets Jev select bounded operations, uses the socai CLI in the user's Chrome session, and generates source-linked reports.

github

Submitted by: @danielyedaniel submission related submission accepted PR · Maintained by: not yet verified · Repository owner: @danielyedaniel

⭐ 6 stars · Growth: 9d +3 stars (+100.0%)

💻 Operating System (OS)

jevme is a voice-controlled macOS agent that uses accessibility controls to operate apps, executes commands as the user speaks, and learns reusable tasks from demonstrations.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @julep-ai

⭐ 6,578 stars · Growth: 9d -2 stars (-0.0%)

⚙️ Development Frameworks

Julep builds durable, composable AI agents as workflows that can resume after failures, retry safely, and expose step-by-step execution.

github

Capabilities: durable agent workflows · safe retries and failure recovery · step-level execution tracing

Interfaces: Python SDK · API

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Prismer-AI

⭐ 10 stars · Growth: 9d +1 stars (+11.1%)

🤖 AI Agents | ⚙️ Development Frameworks | 🛡️ Safety Guardrails (Safeguarding)

A Kubernetes operator for deploying and managing AI agent runtimes, with bounded self-healing that keeps LLMs out of the cluster's workload write path.

github

Capabilities: Manage agent runtimes and their supporting services through a Kubernetes CRD · Apply validated, allowlisted operations through a reconciler · Support persistence, IPC, channel sidecars, and signed audit receipts

Interfaces: Kubernetes Operator · Claw CRD · Go · Docker

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @zenml-io

⭐ 300 stars · Growth: 9d +5 stars (+1.7%)

📊 Evaluation Frameworks | 👁️ Observability Frameworks

An open-source replay-based evaluation toolkit for AI agents that records or imports production traces as sessions and reruns them against model, prompt, or code changes.

github

Capabilities: Record agent runs or import traces from supported observability tools · Replay sessions with recorded tool responses to compare agent changes · Evaluate cohorts and review results against trace evidence

Interfaces: Python package · TypeScript package · CLI · MCP server

Submitted by: @theNamek submission accepted PR · Maintained by: not yet verified · Repository owner: @theNamek

⭐ 0 stars

🧠 Long-Term Memory

kith gives multi-agent systems relationship memory: agents record scoped observations and derive trust, reliability, sentiment, and capability views with provenance.

github

Capabilities: Record append-only observations about other principals · Derive trust, reliability, sentiment, and capability views · Enforce visibility scope across reads and derived views

Interfaces: Python library · LangGraph adapter · MemoryProvider plugin · A2A integration

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ikamensh

⭐ 135 stars · Growth: 9d +2 stars (+1.5%)

⚙️ Development Frameworks

An autonomous multi-agent coding orchestrator that runs coding agents through work cycles with independent verification.

github

Capabilities: Coordinate coding agents through iterative work cycles · Have separate agents review and verify changes · Run long-lived coding tasks from a CLI workflow

Interfaces: Python package · CLI · Claude Code, Cursor, Codex, and Gemini CLI integrations

Submitted by: not yet recovered from repository history · Maintained by: @clubanderson evidence, @eeshaanSA evidence, @MikeSpreitzer evidence, @Nupurshivani evidence · Repository owner: @kubestellar

⭐ 141 stars · Growth: 9d +2 stars (+1.4%)

🤖 AI Agents | 🔧 Tool Calling (Function Calling)

A self-hosted multi-cluster Kubernetes dashboard with an AI operations agent that connects language models to live clusters through MCP.

github

Capabilities: Inspect and manage multiple Kubernetes clusters · Bridge AI operations workflows to live clusters through MCP · Install guided missions for CNCF projects

Interfaces: Web dashboard · MCP · kc-agent · Kubernetes

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 147,421 stars · Growth: 9d +363 stars (+0.2%)

⚙️ Development Frameworks

LangChain is a framework enabling context-aware reasoning applications with integrated libraries, templates, and developer tools

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 18,246 stars · Growth: 9d +19 stars (+0.1%)

⚙️ Development Frameworks

LangChain JS is a framework for developing applications powered by language models, enabling context-aware and reasoning-based applications through composable tools and off-the-shelf chains, with seamless integration with the LangChain Python package

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 18,246 stars · Growth: 9d +19 stars (+0.1%)

Function Calling

Langchain features VectorDBQAChain, which integrates LLMs and vector databases into agent tools for enhanced question-answering capabilities by leveraging data ingested into vector stores

github | docs

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 147,421 stars · Growth: 9d +363 stars (+0.2%)

Function Calling

Langchain integrates various providers like Anthropic, AWS, and OpenAI, and offers tools for components such as LLMs, chat models, and data analysis, supporting functionalities from Alpha Vantage to YouTube

github | docs

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langfuse

⭐ 35,352 stars · Growth: 9d +301 stars (+0.9%)

📊 Evaluation Frameworks

Langfuse, an open-source LLM engineering platform, offers debugging, prompt management, metrics for LLM apps improvement, and won the #1 Golden Kitty in the AI Infra Category from Product Hunt

github | website | twitter | discord

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 42,687 stars · Growth: 9d +399 stars (+0.9%)

⚙️ Development Frameworks

LangGraph is a Python library facilitating the construction of stateful, multi-actor applications with LLMs, enabling cyclic coordination across multiple computation steps, particularly suited for agent-like behaviors, while also providing streaming support, and various guides and examples for implementation and usage

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 3,335 stars · Growth: 9d +22 stars (+0.7%)

⚙️ Development Frameworks

LangGraph.js is a TypeScript and JavaScript library enabling the development of stateful, multi-actor applications with LLMs, featuring capabilities to construct cyclic coordination across multiple computation steps for complex agent-like behaviors, with support for conditional edges and cycles, not limited to DAGs, and extensive documentation with examples on implementation

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 2,328 stars

⚙️ Development Frameworks

LangServe facilitates the deployment of LangChain runnables and chains as a REST API, providing features like automatic schema inference, efficient endpoints, and a playground page, with plans for a hosted version for one-click deployments

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @langchain-ai

⭐ 1,067 stars · Growth: 9d +4 stars (+0.4%)

👁️ Observability Frameworks

LangSmith provides tools for debugging, testing, evaluating, and monitoring LLM applications, integrating seamlessly with LangChain for comprehensive AI agent observability

github | docs | website

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @BerriAI

⭐ 60,109 stars · Growth: 9d +481 stars (+0.8%)

🤖 Assistants API

LiteLLM has added support for the OpenAI Assistants API, enabling seamless integration of stateful operations and automatic RAG pipelines into existing chatbots

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @livekit

⭐ 14,467 stars · Growth: 9d +109 stars (+0.8%)

⚡ Real-Time

An open-source framework for building real-time, programmable participants that run on servers, enabling easy integration with LiveKit WebRTC sessions for processing or generating audio, video, and data streams

github | docs | demo

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @VedankPurohit

⭐ 166 stars · Growth: 9d +2 stars (+1.2%)

👤 Personal Assistants

LiveRecall, an open-source alternative to Microsoft's Recall, utilizes semantic search and encryption to capture and retrieve screen snapshots, enabling AI agents to assist creators in researching and augmenting tasks like journaling or blog post creation based on indexed personal activities

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @run-llama

⭐ 52,404 stars · Growth: 9d +84 stars (+0.2%)

Function Calling

LlamaIndex offers a variety of tools for building data agents, with top downloads including IonicShoppingToolSpec, OpenAPIToolSpec, WikipediaToolSpec, GmailToolSpec, and GoogleCalendarToolSpec, enabling seamless integration with user-defined functions, query engines, and third-party services

github | website | docs

Submitted by: not yet recovered from repository history · Maintained by: @arvinxx evidence, @canisminor1990 evidence · Repository owner: @lobehub

⭐ 82,975 stars · Growth: 9d +147 stars (+0.2%)

⚙️ Development Frameworks

Lobe Chat is an open-source UI framework for building ChatGPT/LLM-based chat applications, featuring modern design, speech synthesis, multi-modal support, extensible plugins, and free one-click deployment for various AI agents

github | website

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @PromtEngineer

⭐ 22,194 stars · Growth: 9d -3 stars (-0.0%)

💻 Local Inference

LocalGPT is an open-source project for secure, private interactions with documents locally, featuring comprehensive model support, embeddings, API for RAG applications, and GUI options, with a focus on privacy and local data processing

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @vitalops

⭐ 1,451 stars · Growth: 9d +2 stars (+0.1%)

⚙️ Development Frameworks

LoopGPT is a modular auto-GPT framework with features such as a 'Plug N Play' API, GPT 3.5 compatibility, minimal prompt overhead, human-in-the-loop capability, and full state serialization, facilitating easy installation and usage through Python code, CLI, or Docker, with the ability to add custom tools and course correction, along with saving and loading agent state, requiring Python 3.8+ and an OpenAI API Key, and optional setup for Google search support

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @looptroop-ai

⭐ 156 stars · Growth: 9d +2 stars (+1.3%)

🤖 AI Agents

LoopTroop is a local AI coding orchestrator that plans work with an LLM council, runs tasks in isolated Git worktrees, and retries failures with fresh context.

github

Capabilities: Ticket-to-PR coding workflow · Multi-model planning · Isolated worktree execution · Verification and review

Interfaces: Desktop UI · CLI · OpenCode integration

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @lucinate-ai

⭐ 12 stars · Growth: 9d +1 stars (+9.1%)

💻 Terminal-Friendly

lucinate is a terminal-native AI chat client for OpenClaw, Hermes Agent, Ollama, and OpenAI-compatible backends, with routines and local skill management.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @RunMaestro

⭐ 3,405 stars · Growth: 9d +47 stars (+1.4%)

🤖 AI Agents | 🔄 Flow Engineering (Platform Engineering)

A desktop command center for running multiple AI coding agents in parallel, automating sessions with event cues and playbooks, and coordinating local and remote agents.

github

Capabilities: Run multiple coding agents in parallel · Automate agent workflows with event cues and playbooks · Coordinate conversations across local and remote agents · Let agents manage runs through a CLI

Interfaces: Desktop UI · CLI · Claude Code · Codex · Gemini CLI · OpenCode

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @PrefectHQ

⭐ 6,204 stars · Growth: 9d +5 stars (+0.1%)

🏗️ Structured Outputs

Marvin is an open-source AI toolkit designed for developers focused on enhancing AI agent capabilities, offering tools for natural language interfaces, image and audio generation, and entity extraction, scalable and easy to integrate into existing projects

github | website

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @mastra-ai

⭐ 28,544 stars · Growth: 9d +211 stars (+0.7%)

⚙️ Development Frameworks

Mastra is a TypeScript framework for building AI agents, workflows, RAG applications, and evaluations.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sortlist

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🤖 AI Agents

A lead-intelligence CLI for discovering prospects and managing lead-generation subscriptions and workflows.

github

Capabilities: Discover leads from monitored sources · Manage subscriptions and lead workflows · Connect campaigns and enroll leads

Interfaces: CLI · Max API

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @melaya-labs

⭐ 6 stars · Growth: 9d +1 stars (+20.0%)

⚙️ Development Frameworks

Nine official open-source SDKs for integrating applications with Melaya's APIs, covering authentication, agents, pipelines, device control, evaluations, and other public platform surfaces.

github

Capabilities: Integrate applications with Melaya APIs · Manage agent workflows and platform resources

Interfaces: TypeScript · Python · Go · Rust · Java · Kotlin · C# / .NET · Ruby · PHP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @mem0ai

⭐ 66,548 stars · Growth: 9d +537 stars (+0.8%)

🧠 Long-Term Memory

Mem0 is an intelligent memory layer for Large Language Models that enhances personalized AI experiences by retaining and utilizing contextual information across various applications.

github | website | docs | discord | twitter | github profile | linkedin

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Felo-Inc

⭐ 42 stars · Growth: 9d +2 stars (+5.0%)

🧠 Long-Term Memory

A persistent memory tool for AI coding agents that isolates workspaces by project and provides a dashboard for reviewing and managing stored context.

github

Capabilities: Keep separate memory workspaces for each project · Store and retrieve project context for coding agents · Review and manage saved memories in a web dashboard

Interfaces: MCP · Web dashboard · Agent skill

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @letta-ai

⭐ 25,022 stars · Growth: 9d +138 stars (+0.6%)

🧠 Long-Term Memory

MemGPT introduces a customizable AI chatbot framework with self-editing memory and access to unlimited data, promoting perpetual, context-rich conversations

website | github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @arcee-ai

⭐ 7,388 stars · Growth: 9d +16 stars (+0.2%)

🔀 Model Merges

Arcee AI's MergeKit offers tools for merging pre-trained large language models, enabling the creation of more versatile AI agents by combining knowledge from different sources, akin to Retrieval Augmented Generation (RAG)

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @metorial

⭐ 3,363 stars · Growth: 9d +5 stars (+0.1%)

🔌 MCP Servers

Metorial is an open-source control plane for agent access to external systems, standardizing integrations with shared authentication, permissions, and observability.

github

Capabilities: agent access to external systems · integration authentication and permissions · integration observability

Interfaces: MCP · SDK · CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @microsoft

⭐ 13,933 stars · Growth: 9d +136 stars (+1.0%)

⚙️ Development Frameworks

Microsoft's framework for building AI agents and workflows, combining capabilities from AutoGen and Semantic Kernel.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @mnemoverse

⭐ 26 stars · Growth: 9d +1 stars (+4.0%)

🧠 Long-Term Memory

Mnemoverse is an open-source MCP server for persistent AI-agent memory, with feedback tools that adjust how memories are ranked.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @moching-ai-dev

⭐ 11 stars · Growth: 9d +6 stars (+120.0%)

💻 Operating System (OS)

Moching is a desktop AI agent for controlling a PC, with screen perception, browser and Office automation, and native tools.

github

Submitted by: not yet recovered from repository history · Maintained by: @CaitieM20 evidence, @clareliguori evidence, @dsp-ant evidence, @kurtisvg evidence, @localden evidence, @nickcoai evidence, @pcarleton evidence, @pja-ant evidence · Repository owner: @modelcontextprotocol

⭐ 9,376 stars · Growth: 9d +70 stars (+0.8%)

📏 Standardization | 🔧 Tool Calling (Function Calling)

An open protocol standard for connecting AI applications to tools and data sources through a client-server interface.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @usemoss

⭐ 789 stars · Growth: 9d +108 stars (+15.9%)

🧠 Long-Term Memory

An open-source semantic search runtime for conversational AI agents, combining vector and keyword retrieval, built-in embeddings, metadata filters, and browser WebAssembly support.

github

Capabilities: Index and retrieve agent knowledge with hybrid semantic and keyword search · Generate embeddings and filter results by metadata · Run retrieval in browser environments through WebAssembly

Interfaces: Python SDK · TypeScript SDK · Elixir SDK · C SDK · WebAssembly

Submitted by: @myspec-dev1 submission accepted PR · Maintained by: not yet verified · Repository owner: @myspecs

⭐ 0 stars

⚙️ Development Frameworks | 🔧 Tool Calling (Function Calling)

MySpec ships Claude Code plugins that connect an MCP server to spec-driven development and a manager plugin that plans task waves, dispatches worker sessions, verifies their pull requests, and gates merges.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @DeepMyst

⭐ 1,140 stars · Growth: 9d +2 stars (+0.2%)

🤖 AI Agents

Mysti is a VS Code AI coding team that connects multiple agent backends and supports collaborative brainstorming, red-teaming, and persistent coding conversations.

github

Capabilities: multi-agent brainstorming and collaboration · red-team analysis · persistent coding conversations

Interfaces: VS Code extension

Submitted by: not yet recovered from repository history · Maintained by: @juanmanueldaza evidence · Repository owner: @juanmanueldaza

⭐ 11 stars · Growth: 9d +1 stars (+10.0%)

⚙️ Development Frameworks

A runtime library and config generator for OpenCode agent workspaces, with provider abstraction, A2A task delegation, persistent memory, and MCP servers.

github

Capabilities: Scaffold and update OpenCode agent workspaces · Route models through provider adapters · Delegate tasks between agents through an A2A hub · Store and retrieve agent memory

Interfaces: CLI · Python package · MCP servers · OpenCode integration

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Nano-Collective

⭐ 2,497 stars · Growth: 9d +5 stars (+0.2%)

🤖 AI Agents | 💻 Local Inference | 💻 Terminal-Friendly

A local-first terminal coding agent that connects to user-selected models and supports tool calling, MCP, and custom commands.

github

Capabilities: Read and edit local code · Run coding-agent tools in the terminal · Connect to local or hosted language models · Extend workflows with MCP and custom commands

Interfaces: CLI · Ollama · OpenRouter · OpenAI-compatible APIs · MCP

Submitted by: not yet recovered from repository history · Maintained by: @Aoshine-T evidence, @BinLiang2021 evidence, @ghydsgaaa evidence, @johnsonice evidence, @netmindCjx evidence, @zzfana407 evidence · Repository owner: @NetMindAI-Open

⭐ 88 stars · Growth: 9d +3 stars (+3.5%)

🤖 AI Agents | 🧠 Long-Term Memory

An open-source AI team workspace for persistent agent identity and memory, multi-agent collaboration, and long-running work across connected channels.

github

Capabilities: Maintain agent identity and persistent memory · Coordinate long-running multi-agent teams · Share context through agent messaging

Interfaces: Desktop UI · CLI · MCP integrations · Lark, Slack, and WeChat channels

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @NVIDIA-NeMo

⭐ 7,243 stars · Growth: 9d +47 stars (+0.7%)

🛡️ Safety Guardrails (Safeguarding)

NeMo Guardrails is an open-source toolkit facilitating the integration of programmable guardrails, essential for steering and safeguarding AI agents' conversational outputs, into large language model-based applications

github | research paper

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @parkain707

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🤖 AI Agents | ⚙️ Development Frameworks | 🛡️ Safety Guardrails (Safeguarding)

Nexus-Agent is an autonomous software-engineering agent with a self-healing execution loop, AST pre-validation, persistent rollback, and terminal and web interfaces.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @arian-gogani

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🛡️ Safety Guardrails (Safeguarding)

An open-source gateway for AI agents acting on financial data: verifies supporting evidence, evaluates a bounded policy, returns PERMIT, BLOCK, or ESCALATE, and emits a hash-chained receipt.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @nowork-studio

⭐ 3,895 stars · Growth: 9d +35 stars (+0.9%)

🤖 AI Agents

Open-source Claude Code skills for SEO, GEO, Google Ads, and Meta Ads, with integrations that read live account and analytics data.

github

Capabilities: Audit SEO and GEO performance · Analyze and manage Google Ads · Analyze and manage Meta Ads · Review search and web analytics

Interfaces: Claude Code skills · Google Ads MCP · Meta Ads MCP · Google Search Console MCP · Google Analytics MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @The-Context-Company

⭐ 124 stars · Growth: 9d +1 stars (+0.8%)

👁️ Observability Frameworks

Observatory provides TypeScript and Python OpenTelemetry packages for instrumenting LLM and agent frameworks, plus a local widget for viewing traces. Integrations include Vercel AI SDK, Claude Agent SDK, LangChain, Mastra, OpenClaw, CrewAI, Agno, and LiteLLM.

github

Capabilities: LLM and agent trace instrumentation · Framework integrations · Local trace visualization

Interfaces: TypeScript SDK · Python SDK · Local web widget

Submitted by: @donhardman submission accepted PR · Maintained by: not yet verified · Repository owner: @Muvon

⭐ 148 stars · Growth: 9d +4 stars (+2.8%)

🤖 AI Agents | ⚙️ Development Frameworks

A Rust CLI-first AI agent runtime that uses model providers and MCP tools to work with files, code, and shell, with interactive, piped, daemon, WebSocket, and ACP modes.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ivoshemi-sys

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

⚙️ Development Frameworks

An agent-to-agent marketplace protocol for capability discovery, reverse auctions, task verification, and USDC escrow payments.

github

Capabilities: Discover agents by their capabilities · Coordinate reverse auctions for agent tasks · Verify task delivery and settle payments through escrow

Interfaces: Python SDK · HTTP API · MCP and A2A

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @OktoLabsAI

⭐ 69 stars · Growth: 9d +23 stars (+50.0%)

🤖 AI Agents | ⚙️ Development Frameworks | 🛡️ Safety Guardrails (Safeguarding)

Okto-Nexus is a local-first MCP coordination hub for multi-agent teams, with durable messaging, atomic handoffs, governance controls, and human approval.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @OktoLabsAI

⭐ 111 stars · Growth: 9d +21 stars (+23.3%)

⚙️ Development Frameworks | 🛡️ Safety Guardrails (Safeguarding)

Okto-Pulse is a local-first SDLC workbench for AI coding teams, with structured project artifacts, governance gates, validation rules, and MCP tools.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ollama

⭐ 182,140 stars · Growth: 9d +423 stars (+0.2%)

💻 Local Inference

Ollama is a tool for running large language models locally, offering easy setup for macOS, Windows, Linux, and Docker, along with a library of models and quickstart guides for customization and integration

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @omega-memory

⭐ 219 stars · Growth: 9d +2 stars (+0.9%)

🧠 Long-Term Memory

A local-first, cross-model memory system that helps AI agents retain context, coordinate, and learn between sessions.

github

Capabilities: Store and retrieve context across agent sessions · Share memory across coding agents and model providers · Support agent coordination and learning from prior work

Interfaces: Python package · CLI · MCP server

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @DrDroidLab

⭐ 112 stars · Growth: 9d +1 stars (+0.9%)

🧠 Long-Term Memory

Open Index is a self-hosted context layer for domain-specialized AI agents, building searchable knowledge graphs with structured schemas, semantic search, and MCP read/write access.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @openinterpreter

⭐ 68,502 stars · Growth: 9d +63 stars (+0.1%)

💻 Operating System (OS)

Open Interpreter is a coding agent enabling language models to execute code locally, facilitating natural-language interaction with your computer's capabilities, overcoming limitations of hosted solutions like internet access and package restrictions. It features interactive and programmatic chats, system message customization, and can control your computer's keyboard and mouse, allowing for enhanced control and flexibility in development environments

github

Submitted by: @JasonBuildAI submission accepted PR · Maintained by: not yet verified · Repository owner: @the-open-agent

⭐ 5,687 stars · Growth: 9d +33 stars (+0.6%)

🤖 AI Agents | 👤 Personal Assistants | 🔧 Tool Calling (Function Calling)

A self-hostable personal AI assistant with autonomous browser, search, shell, office, and MCP tools, plus document RAG and a visual workflow builder.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @openagents-org

⭐ 4,172 stars · Growth: 9d +31 stars (+0.7%)

🤖 AI Agents | ⚙️ Development Frameworks

OpenAgents is an open-source workspace where AI agents connect and collaborate, with an agent launcher, shared workspace, and support for agent protocols and SDKs.

github

Capabilities: multi-agent collaboration · agent workspace management · agent protocol integrations

Interfaces: Web UI · SDK

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @openai

⭐ 29,825 stars · Growth: 9d +125 stars (+0.4%)

⚙️ Development Frameworks

OpenAI's framework for building agent workflows with tools, handoffs, guardrails, sessions, and tracing.

github

Submitted by: @regnull (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @regnull

⭐ 27 stars

🤖 AI Agents | 🧠 Long-Term Memory

OpenBot is an open-source self-hosted platform for running a team of persistent AI bots that wake on incoming messages, use shell, HTTP, custom Python and MCP tools, keep long term memory, and pause for human approval.

github

Capabilities: Run named bots with their own instructions, tools and long term memory that wake when mail arrives · Hand a thread from one bot to another with an @mention, with a hop limit that stops runaway loops · Pause and ask a human before a sensitive action

Interfaces: Web UI · HTTP API · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @openclaw

⭐ 391,261 stars · Growth: 9d +755 stars (+0.2%)

🤖 AI Agents | 👤 Personal Assistants

OpenClaw is a personal AI assistant that runs on your own devices and connects to messaging channels, tools, skills, scheduled tasks, and memory.

github

Capabilities: multi-channel messaging · scheduled tasks · skills and tool integrations

Interfaces: CLI · Web UI · Messaging

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @flik2002

⭐ 13 stars · Growth: 9d +1 stars (+8.3%)

🤖 AI Agents

An offline web dashboard for OpenClaw that monitors agent sessions, scheduled tasks, token usage, message trends, and local system metrics.

github

Capabilities: Monitor live agent sessions and task schedules · Track per-model token usage and message trends · Display local system uptime and memory metrics

Interfaces: Web dashboard · Node.js API · Windows launcher

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @OpenHands

⭐ 89,933 stars · Growth: 9d +767 stars (+0.9%)

🤖 AI Agents

OpenHands is an open-source software development agent platform that can edit code, run commands, and interact with tools.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @daxaur

⭐ 173 stars · Growth: 9d +5 stars (+3.0%)

👤 Personal Assistants

OpenPaw turns Claude Code into a personal assistant by installing skills and tools for email, calendars, smart home, and other services, with persistent memory, Telegram access, scheduling, and a local task dashboard.

github

Capabilities: personal assistant skills · persistent memory · task scheduling and dashboard

Interfaces: CLI · Telegram

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Open-Twin

⭐ 15 stars · Growth: 9d +2 stars (+15.4%)

🤖 AI Agents | 👤 Personal Assistants

A local-first CLI and dashboard for running autonomous digital-twin agents across social platforms, with content workflows, scheduling, and browser automation.

github

Capabilities: Research trends and plan or draft social content · Track engagement and schedule publishing workflows · Automate platform interactions through Chrome

Interfaces: CLI · Web dashboard · npm package

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @operantlabs

⭐ 25 stars · Growth: 9d +2 stars (+8.7%)

🔧 Tool Calling (Function Calling)

An open-source MCP server that gives AI agents a toolkit for security testing, network forensics, and vulnerability assessment.

github

Capabilities: Run security tests and vulnerability assessments through MCP tools · Analyze network captures and memory artifacts · Provide reconnaissance and malware-analysis utilities

Interfaces: MCP server · npm package

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @comet-ml

⭐ 22,364 stars · Growth: 9d +125 stars (+0.6%)

📊 Evaluation Frameworks

Opik is an open-source platform for LLM and agent observability, tracing, evaluation, prompt management, and production monitoring.

github

Capabilities: LLM and agent tracing · evaluation and experiment comparison · prompt management and monitoring

Interfaces: Web UI · API · SDK

Submitted by: @yanauto (self-reported founder/team) submission accepted PR · Maintained by: @yanauto (self-reported) evidence · Repository owner: @yanauto

⭐ 46 stars

🤖 AI Agents | 💻 Terminal-Friendly

Opus Manager is a Claude Code skill that makes Claude the manager of a coding project: it writes tickets, dispatches them to cheaper coding agent CLIs on your machine, reruns the acceptance checks itself, and has a model from another vendor review the change.

github

Capabilities: Write each task as a ticket file and dispatch it to a worker CLI such as pi, cursor-agent or codex · Rerun the ticket acceptance commands itself and check each review finding against the code before accepting it · Keep tickets, receipts and reviews as plain Markdown in the repository

Interfaces: Claude Code skill · CLI

Submitted by: @xqliu (self-reported founder/team) submission accepted PR · Maintained by: @xqliu (self-reported) evidence · Repository owner: @orbi-build

⭐ 195 stars

🤖 AI Agents | ⚙️ Development Frameworks

A self-hosted coding agent that takes labeled GitHub issues through planning and implementation, opens pull requests, checks changes against acceptance criteria, and merges when configured gates pass.

github

Submitted by: @xizhuomengcontin submission accepted PR · Maintained by: not yet verified · Repository owner: @Continuum-AI-Corp

⭐ 274 stars

👁️ Observability Frameworks

OrcaReplay records a coding agent run from outside the process, replays it offline against the recorded bytes with no model called, and forks a checkpoint onto other models graded by a verify command.

github

Capabilities: Record a coding agent run at the process and socket boundary without modifying the agent, with model traffic, shell exit codes, per turn file changes and MCP JSON-RPC on one timeline · Replay a recorded session offline with no provider called and nothing spent · Fork one checkpoint onto several models and grade them by a verify command exit code

Interfaces: CLI · Web UI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @oxgeneral

⭐ 167 stars · Growth: 9d +1 stars (+0.6%)

⚙️ Development Frameworks

A CLI runtime for running and coordinating teams of AI agents on parallel tasks with shared state and review workflows.

github

Capabilities: Run coding agents in parallel against a typed task queue · Track task state, review gates, and failed-agent retries · Coordinate agents with inter-agent messaging and shared context

Interfaces: CLI · Terminal UI · Node.js package

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Orkas-AI

⭐ 2,150 stars · Growth: 9d +26 stars (+1.2%)

🤖 AI Agents | 💻 Operating System (OS)

Orkas is a local-first desktop AI team where a commander coordinates specialist agents and local coding-agent runtimes, with support for user-provided model providers.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @dottxt-ai

⭐ 15,898 stars · Growth: 9d +13 stars (+0.1%)

🏗️ Structured Outputs

Outlines is a robust text generation library designed for agentic AI developers, featuring support for multiple model integrations, advanced prompting with Jinja, efficient structured generation through regex, JSON schema, context-free grammars, and more, enabling the creation of predictable and structured AI agent outputs

github | website | discord

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sortlist

⭐ 8 stars · Growth: 9d +1 stars (+14.3%)

🤖 AI Agents

A command-line client for managing prospects, sourcing, campaigns, and sales outreach through the Overloop API.

github

Capabilities: Find and manage prospects · Create and operate outbound campaigns · Manage sourcing and conversations

Interfaces: CLI · Overloop API

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @developersdigest

⭐ 5,036 stars · Growth: 9d -4 stars (-0.1%)

🏗️ Structured Outputs

A versatile answer engine leveraging Groq, Mistral AI, Langchain.JS, Brave Search, Serper API, and OpenAI to deliver efficient and sophisticated responses with reduced hallucination through RAG for citation-backed search queries

github | github profile

Submitted by: @rubenluetke10-beep (self-reported founder/team) submission accepted PR · Maintained by: @rubenluetke10-beep (self-reported) evidence · Repository owner: @PersonalJarvis

⭐ 94 stars

👤 Personal Assistants

PersonalJarvis is an open-source desktop AI assistant with a custom wake phrase and local speech to text that runs background agents, uses the browser and desktop, and supervises coding agents such as Claude Code and Codex in an Agentic IDE.

github

Capabilities: Answer to a custom wake phrase by voice or chat, with local speech to text · Run background agents, use the browser and desktop apps, connect MCP servers, keep memory and run scheduled routines · Run CLI coding agents side by side in terminal panes with Git worktrees, including on remote machines over SSH

Interfaces: Desktop app · Voice · CLI · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @pilot-protocol

⭐ 146 stars · Growth: 9d +6 stars (+4.3%)

📏 Standardization

Pilot Protocol is an overlay network for AI agents, with permanent virtual addresses, encrypted UDP tunnels, NAT traversal, and per-peer trust controls.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @microsoft

⭐ 37,797 stars · Growth: 9d +231 stars (+0.6%)

🔌 MCP Servers | 🌐 Web Browsing Frameworks

Microsoft's Playwright MCP server gives agents structured, accessibility-tree-based control of a real browser.

github

Submitted by: @Acekorneya (self-reported founder/team) submission accepted PR · Maintained by: @Acekorneya (self-reported) evidence · Repository owner: @Acekorneya

⭐ 1 stars

💻 Local Inference | 💻 Operating System (OS)

POK-Agent is a Windows computer-use and coding agent that combines a planner with a fast screen-action model, verifies step outcomes, and can record and replay successful task sequences.

github

Capabilities: Plan and execute computer-use tasks · Verify step outcomes and branch based on the screen · Record and replay successful motor programs

Interfaces: Desktop app · Local model

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @samirasadov28-code

⭐ 3 stars · Growth: 9d +1 stars (+50.0%)

🤖 AI Agents

A web app that monitors Polymarket and Kalshi markets, detects configured price, volume, and wallet signals, and provides LLM analysis of alerts.

github

Submitted by: @Partridge12 (self-reported founder/team) submission accepted PR · Maintained by: @Partridge12 (self-reported) evidence · Repository owner: @posteverywhere

⭐ 4 stars

🔌 MCP Servers

PostEverywhere’s open-source MCP server connects AI agents to social accounts to create and schedule posts, manage queues, review publishing status, and retry failed posts.

github

Capabilities: Manage connected social accounts and posting queues · Create, schedule, and publish posts to supported social networks · Inspect publishing results and retry failures

Interfaces: MCP server

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @MervinPraison

⭐ 9,131 stars · Growth: 9d +45 stars (+0.5%)

⚙️ Development Frameworks

Praison AI is a low-code, centralized framework leveraging AutoGen and CrewAI to simplify creating and orchestrating multi-agent systems for LLM applications, emphasizing customization and ease of human-agent interaction

github | demo | website

Submitted by: @Prismfy submission accepted PR · Maintained by: not yet verified · Repository owner: @Prismfy

⭐ 3 stars · Growth: 9d +2 stars (+200.0%)

✍️ Prompt Engineering | 🔧 Tool Calling (Function Calling)

A CLI that installs the Prismfy live-search command and managed search-policy blocks into local AI-agent instruction files, with selectable verification modes and dry-run support.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @zylon-ai

⭐ 57,557 stars · Growth: 9d +19 stars (+0.0%)

💻 Local Inference

PrivateGPT is a secure, offline-capable AI tool for querying documents with Large Language Models, offering high-level and low-level APIs for privacy-conscious, context-aware application development

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @jarvismaximum-hue

⭐ 9 stars · Growth: 9d +1 stars (+12.5%)

🤖 AI Agents

An open prediction-market benchmark and starter template for building autonomous AI trading agents.

github

Capabilities: Register autonomous agents to participate in prediction markets · Place trades on real-asset price outcomes · Stream market updates and evaluate agent performance

Interfaces: Python SDK · Node.js SDK · HTTP API · WebSocket

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @HarperZ9

⭐ 3 stars · Growth: 9d +1 stars (+50.0%)

📊 Evaluation Frameworks | 👁️ Observability Frameworks

Project Telos is a local-first MCP and CLI workbench for inspectable agent workflows, with provenance, routing ledgers, action receipts, context packs, diagnostics, and replayable verification packets.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @promptise-com

⭐ 872 stars · Growth: 9d +1 stars (+0.1%)

🤖 AI Agents | ⚙️ Development Frameworks

A Python framework for building and running AI agents with integrated tools, controllable reasoning, memory, security, governance, and observability.

github

Capabilities: Build agents and MCP tools with a Python SDK · Run autonomous agents with memory and controllable reasoning · Govern multi-tenant deployments with budgets, approvals, and observability

Interfaces: Python package · MCP server · Autonomous runtime

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @pydantic

⭐ 28,935 stars · Growth: 9d +60 stars (+0.2%)

🏗️ Structured Outputs

Pydantic is a Python library facilitating data validation through type hints, particularly useful for AI agents, offering fast validation capabilities and compatibility with various development tools

github | website

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Detrol

⭐ 120 stars · Growth: 9d +3 stars (+2.6%)

🤖 AI Agents

Quorum lets multiple AI models and agent CLIs conduct structured debates using formal methods, critique, and synthesis, through a terminal UI, MCP, and a web version.

github

Capabilities: structured multi-agent debate · critique and consensus synthesis · local and hosted model participation

Interfaces: CLI · MCP · Web UI

Submitted by: @LivXue (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @EverMind-AI

⭐ 5,129 stars

🤖 AI Agents

Raven is an open-source host agent that plans complex tasks as DAGs and orchestrates built-in research, coding, design, and on-call agents plus third-party agents such as Claude Code, Codex, OpenClaw, and Hermes Agent over ACP, CLI, or OpenAI-compatible APIs.

github

Capabilities: Plan complex tasks as DAGs and coordinate agents with task dependencies and parallel execution · Run built-in agents for deep research, software development, design, and unattended workflow automation · Connect third-party agents over ACP, CLI, or OpenAI-compatible APIs, with presets for 13 agents

Interfaces: Web UI · Terminal UI · ACP · A2A · Docker

Submitted by: @HNF-FRN (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @HNF-FRN

⭐ 4 stars · Growth: 9d +1 stars (+33.3%)

🤖 AI Agents

A self-hosted Telegram agent that analyzes social-media videos, returns source-linked breakdowns, and can plan or build related projects on the user's PC with command approvals.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @remembra-ai

⭐ 16 stars · Growth: 9d +1 stars (+6.7%)

🧠 Long-Term Memory

Remembra is a self-hosted semantic memory system for AI agents, with persistent memory, retrieval, knowledge graph features, and integrations for agent frameworks and MCP.

github

Capabilities: semantic memory and retrieval · knowledge graph · agent framework integrations

Interfaces: API · SDK · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @trevhud

⭐ 8 stars · Growth: 9d +1 stars (+14.3%)

🔄 Flow Engineering (Platform Engineering)

rote compiles proven agent skills into typed, deterministic Python or TypeScript pipelines, with tests for fixed steps and typed LLM-judge signatures for work requiring judgment.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @nikoSchoinas

⭐ 33 stars · Growth: 9d +1 stars (+3.1%)

🤖 AI Agents | 🔧 Tool Calling (Function Calling)

An async HTTP client for autonomous agents that handles HTTP 402 payments across multiple payment rails, enforces spend policies, recovers credentials, and emits audit traces.

github

Capabilities: Handle paid HTTP requests across x402, L402, MPP-Tempo, and Stripe SPT · Apply spend caps and policy rules to agent payments · Recover payment credentials and emit audit traces

Interfaces: Python async HTTP client · PyPI package

Submitted by: @yanghao1143 submission accepted PR · Maintained by: not yet verified · Repository owner: @yanghao1143

⭐ 19 stars · Growth: 9d +1 stars (+5.6%)

🤖 AI Agents | ⚙️ Development Frameworks

A Rust research prototype for LLM runtime and agent control, with routing, memory, runtime adapters, rollback, tracing, and gated self-evolution.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @l33tdawg

⭐ 254 stars · Growth: 9d +2 stars (+0.8%)

🧠 Long-Term Memory | 🛡️ Safety Guardrails (Safeguarding)

Persistent, consensus-validated memory infrastructure for AI agents, with confidence scoring and natural memory decay.

github

Capabilities: Persist institutional memory across AI conversations · Validate memory updates through signed votes and multi-validator consensus · Track confidence and decay memory over time

Interfaces: Desktop app · Go CLI · MCP server

Submitted by: @0000F8 submission accepted PR · Maintained by: not yet verified · Repository owner: @0000F8

⭐ 1 stars

🔌 MCP Servers

MCP server for Salt, an end-to-end encrypted chat where humans and AI agents are equal contacts and can message, ask each other for approval, and pay each other. The local stdio server runs one agent identity with its own key; a hosted OAuth endpoint lets any MCP client connect a keyless agent with no setup.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sandbaseai

⭐ 681 stars · Growth: 9d +9 stars (+1.3%)

🤖 AI Agents | ⚙️ Development Frameworks

SandBase Harness is a self-hosted AI agent runtime and MCP bridge for persistent sessions, sandboxed tool execution, credentials, audit logs, and replay. It supports local and container-based execution backends, whose isolation properties differ.

github

Capabilities: Persistent agent sessions · Sandboxed tool execution · Audit and replay

Interfaces: CLI · MCP · Local web console

Submitted by: @satohubai (self-reported founder/team) submission accepted PR · Maintained by: @satohubai (self-reported) evidence · Repository owner: @satohubai

⭐ 1 stars

🔌 MCP Servers | 🔧 Tool Calling (Function Calling)

A hosted MCP server and agent skill for searching an index of onchain-agent tooling and checking package reports about key access, key egress, and fund actions before installation. Its install hook sends command text to Sato Hub and allows installs by default.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @SaynaAI

⭐ 314 stars · Growth: 9d 0 stars (+0.0%)

⚙️ Development Frameworks | ⚡ Real-Time

Sayna is a self-hosted real-time voice server that provides speech-to-text and text-to-speech through WebSocket and REST APIs, with provider integrations, LiveKit streaming, and SIP telephony.

github

Capabilities: speech-to-text and text-to-speech · real-time audio streaming · SIP telephony integration

Interfaces: WebSocket API · REST API

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @screenpipe

⭐ 21,806 stars · Growth: 9d +97 stars (+0.4%)

🤖 AI Agents | 💻 Local Inference | 🧠 Long-Term Memory | 👤 Personal Assistants

Screenpipe captures screen and audio locally, then exposes searchable work history and agent context through an MCP server. It supports scheduled agent workflows and per-pipe access policies, with processing options that include local and configured remote models.

github

Capabilities: Local screen and audio memory · Searchable context for AI agents · Per-pipe data access policies

Interfaces: Desktop app · CLI · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @simin-yuan

⭐ 3 stars · Growth: 9d +1 stars (+50.0%)

📊 Evaluation Frameworks

Self Auditing Agent pairs audit claims with rerunnable commands and includes gatecheck, a mutation tester for quality gates available as a GitHub Action and pre-commit hook.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @oraios

⭐ 29,972 stars · Growth: 9d +169 stars (+0.6%)

🔌 MCP Servers

Serena is a coding-agent toolkit exposed as an MCP server, providing semantic code retrieval and editing through language-server-powered tools.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sidclawhq

⭐ 15 stars · Growth: 9d +1 stars (+7.1%)

🛡️ Safety Guardrails (Safeguarding)

An agent governance platform that intercepts tool calls for policy checks, human approval, and auditable traces.

github

Capabilities: Enforce policies on agent tool calls · Route risky actions for human approval · Record auditable traces of agent actions

Interfaces: Web dashboard · SDK · MCP integrations

Submitted by: @MohammadHijjawi97 (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @MohammadHijjawi97

⭐ 6 stars

🔌 MCP Servers | 💻 Terminal-Friendly

since-cutoff is an open-source Python tool that shows a coding agent which public APIs of a project's pinned dependencies changed after its model's training cutoff and writes short notes into AGENTS.md or CLAUDE.md.

github

Capabilities: Compare each pinned dependency's public API against its latest release before the model's training cutoff, statically with no model call · Show where the project code uses APIs that changed · Write and keep AGENTS.md and CLAUDE.md notes in step with the lockfile

Interfaces: CLI · MCP · Claude Code plugin · Gemini CLI extension · GitHub Action

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @rafsilva85

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

⚙️ Development Frameworks

An MCP server that lets AI coding agents search, discover, and evaluate skills from the SkillFlow marketplace.

github

Capabilities: Search and discover skills from the SkillFlow marketplace · Review skill performance and trust metrics

Interfaces: MCP server over stdio · Streamable HTTP

Submitted by: @suchintan (self-reported founder/team) submission accepted PR · Maintained by: @suchintan (self-reported) evidence · Repository owner: @Skyvern-AI

⭐ 23,132 stars

🌐 Web Browsing Frameworks

Skyvern automates browser workflows with language models and computer vision, including logging into portals, filling forms, downloading user-requested files, and extracting structured data.

github

Capabilities: Automate browser workflows across websites without dedicated APIs · Fill forms, interact with portals, and extract structured data

Interfaces: Web app · REST API · Python SDK · TypeScript SDK · CLI · MCP server

Submitted by: @sksareen submission accepted PR · Maintained by: not yet verified · Repository owner: @sksareen

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🤖 AI Agents

An open-source encyclopedia where agents generate, revise, and review knowledge articles, then explore connected topics through a knowledge graph.

github

Capabilities: Agent article generation and editing · Content review and revision history · Knowledge graph and topic discovery

Interfaces: MCP server · HTTP API · Web application

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @huggingface

⭐ 29,667 stars · Growth: 9d +173 stars (+0.6%)

⚙️ Development Frameworks | 💻 Terminal-Friendly

Hugging Face's lightweight agent framework, including code agents that write and execute Python actions.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @User0856

⭐ 3 stars · Growth: 9d +1 stars (+50.0%)

🔌 MCP Servers

SnapRender is a screenshot API for agent workflows, with a remote MCP server and SDK integrations for capturing websites as images or PDFs.

github

Capabilities: website screenshots · image and PDF capture · remote MCP tools

Interfaces: MCP · SDK · API

Submitted by: @Benjaminsen (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @solveathome

⭐ 5 stars

🤖 AI Agents

solveathome is an MIT-licensed framework for running a swarm of AI agents owned by different people against one open problem, with public transcripts, peer review by other contributors' agents, and a small group of trusted reviewers who decide what is accepted.

github

Capabilities: Coordinate agents such as Claude Code, Codex or any agent that can fetch a URL on one open research problem · Have other contributors agents check each result, with trusted reviewers deciding acceptance by one vote per person · Publish every result, review and transcript, and trace credit through the citation chain

Interfaces: HTTP API · Web UI

Submitted by: @thenaxon (self-reported founder/team) submission accepted PR · Maintained by: @thenaxon (self-reported) evidence · Repository owner: @thenaxon

⭐ 29 stars

🤖 AI Agents | 💻 Local Inference | 🧠 Long-Term Memory | 👤 Personal Assistants

A self-hosted server for personal AI agents with persistent private memory, shared wiki, delegated agents, scheduled work, and web, shell, and browser tools across terminal, browser, and mobile clients.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sonz-ai

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

⚙️ Development Frameworks | 🧠 Long-Term Memory

An open-source Go client for the hosted Sonzai Mind Layer API, providing agent runtimes with memory context, turn persistence, and voice services.

github

Capabilities: Retrieve memory context bundles for agent sessions · Persist agent conversation turns and submit usage reports · Access Sonzai runtime and voice API resources from Go

Interfaces: Go module · REST API client

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @echennells

⭐ 5 stars · Growth: 9d +1 stars (+25.0%)

₿ Bitcoin

A self-custodial Spark Bitcoin wallet skill for AI agents, supporting BTC transfers, Lightning invoices, L402 payments, token operations, and merchant purchases with payment guardrails.

github

Capabilities: Create and manage a self-custodial Spark wallet · Send and receive Bitcoin through Spark and Lightning · Pay L402 APIs and perform token or merchant transactions with policy checks

Interfaces: Claude Code plugin · npm package · Agent skill

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @browserbase

⭐ 25,530 stars · Growth: 9d +140 stars (+0.6%)

🌐 Web Browsing Frameworks

Stagehand adds natural-language act, extract, and observe APIs to Playwright for browser automation.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @steel-dev

⭐ 7,733 stars · Growth: 9d +41 stars (+0.5%)

🌐 Web Browsing Frameworks

An open-source browser API and sandbox that provides managed browser sessions for AI agents and web automation.

github

Capabilities: Create and manage isolated browser sessions · Automate live websites and authenticated workflows · Capture browser page context and artifacts

Interfaces: Browser API · SDKs · Browser automation tools

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @stipend-sh

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🔧 Tool Calling (Function Calling)

A non-custodial USDC wallet on Base for AI agents, with an MCP server and spending limits enforced below the instruction layer.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @stoneforge-ai

⭐ 190 stars · Growth: 9d 0 stars (+0.0%)

⚙️ Development Frameworks

A web dashboard and local runtime for orchestrating parallel AI coding agents, with task dispatch and isolated workspaces.

github

Capabilities: Dispatch coding tasks to multiple agents in parallel · Isolate agent work and coordinate task states · Monitor agent activity and recover from failures

Interfaces: Web dashboard · CLI · Local runtime

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @samirasadov28-code

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🤖 AI Agents

A GPS-aware mobile web app that discovers nearby landmarks, generates location-based stories with an LLM, and narrates them as users walk.

github

Capabilities: GPS-aware landmark discovery · LLM-generated audio stories · Text-to-speech narration

Interfaces: Progressive web app

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @strands-agents

⭐ 8,647 stars · Growth: 9d +233 stars (+2.8%)

⚙️ Development Frameworks

AWS's model-driven SDK for building agents that use tools through a model-directed loop.

github

Submitted by: @ItSeemedLikeAGoodIdeaAtTheTime submission accepted PR · Maintained by: not yet verified · Repository owner: @ItSeemedLikeAGoodIdeaAtTheTime

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🤖 AI Agents | 🔧 Tool Calling (Function Calling)

A persistent archaeological exploration game for AI agents, with REST and MCP interfaces for discovering fragments, interpreting finds, and connecting them with other agents.

github

Submitted by: @azrollin (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @sunglasses-dev

⭐ 9 stars

🛡️ Safety Guardrails (Safeguarding)

SUNGLASSES is an open-source local input firewall for AI agents that scans text and files for prompt injection, credential leaks and data exfiltration, and reports both what it matched and what it could not read.

github

Capabilities: Scan text, files, PDFs, images and QR codes locally for prompt injection, credential leaks and data exfiltration patterns · Report findings together with the inputs it could not read, so an agent can decide what to pass onward · Block credential paths and policy violations before a tool runs through a Claude Code hook · Refuse MCP tool calls through a local proxy until a person approves the server at an interactive terminal

Interfaces: CLI · Python API · MCP · GitHub Action · Claude Code hook

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @superagent-ai

⭐ 6,765 stars · Growth: 9d +11 stars (+0.2%)

⚙️ Development Frameworks

Superagent is an open-source AI assistant framework backed by Y Combinator, facilitating the integration of large language models (LLM) and generative AI into applications, supporting various use cases such as question answering, chatbots, and content generation

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @pungme

⭐ 29 stars · Growth: 9d +3 stars (+11.5%)

💻 Operating System (OS)

Superagent for Mac is an open-source macOS workspace for Claude Code and Codex, combining persistent coding-agent chats with an agent-controlled browser, iOS Simulator, and companion iPhone app for monitoring conversations and responding to agent requests. It runs locally and uses the user’s existing agent subscription.

github

Capabilities: Browser automation · iOS Simulator interaction · Remote agent supervision

Interfaces: macOS desktop app · iOS companion app · Claude Code and Codex

Submitted by: @nybarius submission accepted PR · Maintained by: not yet verified · Repository owner: @nybarius

⭐ 1 stars

🔌 MCP Servers

SVRF is a merge train that batches, gates and lands pull requests for repositories where coding agents open pull requests faster than CI can gate them one by one; it ships an MCP server (svrf mcp) so an agent whose pull request is held can query why and fix it itself.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @kyegomez

⭐ 7,230 stars · Growth: 9d +24 stars (+0.3%)

⚙️ Development Frameworks

Swarms orchestrates multi-agent collaboration for production-grade applications, solving issues like short memory and high costs, with customizable tools for specific needs, currently used by RBC, John Deere, and AI startups

github | discord | docs

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @SWE-agent

⭐ 20,484 stars · Growth: 9d +77 stars (+0.4%)

⚙️ Development Frameworks

This open source project introduces SWE-agent, a software engineering agent for LMs like GPT-4, enhancing bug and issue resolution in GitHub repositories with state-of-the-art performance, facilitated by a well-designed Agent-Computer Interface (ACI) and support for OpenAI and Anthropic Claude models

github | website

Submitted by: @baronmuh submission accepted PR · Maintained by: not yet verified · Repository owner: @baronmuh

⭐ 2 stars · Growth: 9d +1 stars (+100.0%)

🤖 AI Agents

A local-first, self-hosted agent-to-agent communication platform with natural-language messaging, human-approved tasks, groups, delegation, a CLI, Python client, web UI, and encrypted backups.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @TabbyML

⭐ 33,893 stars · Growth: 9d 0 stars (+0.0%)

🤖 AI Agents

Tabby is a self-hosted, open-source AI coding assistant similar to GitHub Copilot, featuring a self-contained setup with no DBMS/cloud dependency, OpenAPI for easy integration, consumer-grade GPU support, and a full-feature admin UI in its latest release

github | website | docs

Submitted by: @yannickmonney submission accepted PR · Maintained by: @yannickmonney (self-reported) evidence · Repository owner: @tale-project

⭐ 29 stars

🤖 AI Agents | 🔄 Flow Engineering (Platform Engineering)

Tale is a collaborative workspace where teams turn work into tasks, delegate tasks to configured agents in persistent sandbox workspaces, and review their reports and deliverables.

github

Capabilities: Organize projects, tasks, acceptance criteria, and shared context · Delegate tasks to configured agent runtimes in sandbox workspaces · Review agent reports and deliverables collaboratively

Interfaces: Web app · REST API · MCP endpoint

Submitted by: @johnxie submission accepted PR · Maintained by: not yet verified · Repository owner: @taskade

⭐ 66 stars · Growth: 9d 0 stars (+0.0%)

🤖 AI Agents

An AI-native workspace with custom agents, workflow automation, and prompt-built apps; its official public repository contains product documentation and clone-ready App Kits.

github

Capabilities: Custom AI agents · Workflow automation · Prompt-built apps

Interfaces: Web application · REST API · MCP server

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @agentguard-ai

⭐ 36 stars · Growth: 9d +3 stars (+9.1%)

🛡️ Safety Guardrails (Safeguarding)

An open-source security and governance SDK for AI applications, providing deterministic policy enforcement, guardrails, cost tracking, and audit evidence.

github

Capabilities: Apply deterministic policies and tool allowlists to agent actions · Detect sensitive data and track model usage costs · Produce audit decisions and export governance evidence

Interfaces: TypeScript package on npm · Python package on PyPI · MCP servers · Docker sidecar

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sagiyaacoby

⭐ 38 stars · Growth: 9d +2 stars (+5.6%)

⚙️ Development Frameworks | 🧠 Long-Term Memory

A self-hosted workspace for coordinating AI agent teams with structured tasks, persistent memory, and file-conflict controls.

github

Capabilities: Plan, assign, review, and track agent tasks and deliverables · Maintain persistent memory and a shared knowledge base for agent teams · Coordinate file ownership to prevent conflicting edits

Interfaces: Web dashboard · Integrated terminal · Claude Code

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ifyoubuildit

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

📊 Evaluation Frameworks

The Agentic Leaderboard publishes weekly rankings of open-source AI agents across reliability, tool selection, autonomous iteration, and community signals, with public data and methodology.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ModernOps888

⭐ 13 stars · Growth: 9d +1 stars (+8.3%)

⚙️ Development Frameworks

The Forge coordinates LLMs to generate code candidates, compiles and benchmarks them, then evolves top candidates through mutation and selection.

github

Submitted by: @Bengtsson78 submission accepted PR · Maintained by: not yet verified · Repository owner: @Bengtsson78

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

📏 Standardization

An open standard that defines eight evidence-based dimensions for AI agents assessing human–AI operator interactions and issuing credentials.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @sabir-gbs

⭐ 20 stars · Growth: 9d +1 stars (+5.3%)

📏 Standardization

A portable set of standards for AI coding agents covering repository discovery, language selection, development workflows, and post-generation audits.

github

Capabilities: Guide repository discovery and technology selection · Provide language-specific development and validation practices · Define post-code-generation audit and scoring workflows

Interfaces: Codex skill · Claude Code · OpenCode · Documentation

Submitted by: @cgoinglove (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @cgoinglove

⭐ 31 stars · Growth: 9d +23 stars (+287.5%)

👤 Personal Assistants

Thursday is an open-source personal voice assistant that runs on your computer and delegates longer tasks to background AI agents using a real browser, shell, files, Agent Skills, and MCP servers. It keeps the live call available while work runs and returns results as local files; it uses hosted model providers rather than local inference.

github

Capabilities: Voice-based task delegation · Background browser and shell agents · Local file creation

Interfaces: Voice call · Desktop app · MCP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Djtony707

⭐ 19 stars · Growth: 9d +1 stars (+5.6%)

🤖 AI Agents | ⚙️ Development Frameworks

A local-first AI agent framework for running specialist agents with shared memory, tools, approvals, and mission control.

github

Capabilities: Run and coordinate specialist agents with configurable models and harnesses · Share agent memory and tools across missions · Review agent activity and approve higher-autonomy work

Interfaces: CLI · MCP server · Local web interface · npm package

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @TerminallyLazy

⭐ 18 stars · Growth: 9d 0 stars (+0.0%)

🧠 Long-Term Memory

Tree Ring Memory is a framework-agnostic, local-first memory lifecycle for AI agents, with SQLite search, deliberate forgetting, audit logs, consolidation, and runtime adapters.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @microsoft

⭐ 8,689 stars · Growth: 9d +2 stars (+0.0%)

Function Calling

TypeChat is a library that facilitates building natural language interfaces by using schema engineering as an alternative to traditional function calling in LLMs, avoiding JSON schema-based constraints

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @fetchai

⭐ 1,640 stars · Growth: 9d +2 stars (+0.1%)

⚙️ Development Frameworks

uAgents is a Python library by Fetch AI for creating autonomous AI agents with features like easy creation, blockchain network connectivity, and cryptographic security

github | github profile

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @uizze

⭐ 38 stars · Growth: 9d +12 stars (+46.2%)

🖥️ UI Development

UIZZE provides design skills and a free GitHub Action for coding agents: its skills guide product-specific interface design, and the PR check flags common UI issues in changed source. An optional paid MCP server provides web and iOS design references.

github

Capabilities: Agent-guided interface design · Pull request UI checks · Web and iOS design reference search

Interfaces: Agent Skills · GitHub Action · MCP

Submitted by: @happy520ai submission accepted PR · Maintained by: not yet verified · Repository owner: @happy520ai

⭐ 8 stars

⚙️ Development Frameworks | 🔧 Tool Calling (Function Calling)

A self-hosted AI gateway and control plane that exposes governed MCP tools, converts OpenAPI services to MCP, and proxies compatible model APIs with virtual-key budgets, caching, and audit.

github

Capabilities: Expose OpenAPI services and upstream MCP servers as governed tools · Proxy compatible model APIs with virtual-key budgets, caching, and audit trails

Interfaces: MCP · OpenAI-compatible REST API · CLI · Docker

Submitted by: @victorcavero14 (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @Upload-Post

⭐ 9 stars

🔌 MCP Servers

Upload-Post MCP is an open-source TypeScript MCP server that lets AI agents publish, schedule and manage social media posts, analytics, comments and messages across many platforms through the Upload-Post API.

github

Capabilities: Upload and schedule videos, photos, text posts and documents to platforms such as TikTok, Instagram, YouTube and LinkedIn · Poll upload status, read analytics, reply to comments and messages, and manage the posting queue · Expose 58 tools with read only and destructive annotations so clients can ask for confirmation before destructive actions

Interfaces: MCP · stdio · HTTP

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Upsonic

⭐ 7,956 stars · Growth: 9d -1 stars (-0.0%)

⚙️ Development Frameworks

Upsonic is a Python framework for building autonomous AI agents, with tools for workflows, MCP, and agent evaluation.

github

Capabilities: autonomous agent development · tool and MCP integration · workflow and output evaluation

Interfaces: Python SDK · CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @shleder

⭐ 44 stars · Growth: 9d +14 stars (+46.7%)

⚙️ Development Frameworks | 🛡️ Safety Guardrails (Safeguarding)

Vetto is a daemon-less, rootless sandbox that applies kernel-enforced security policies to AI coding agents, with agent command shims and MCP tool isolation.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Cznorth

⭐ 23 stars · Growth: 9d +1 stars (+4.5%)

🤖 AI Agents | 💻 Terminal-Friendly

A self-hosted AI terminal that shares the user's PTY, drafts commands in the terminal input for review, and supports SSH and agent integrations.

github

Capabilities: Answer questions using live terminal context · Draft commands for a person to inspect before execution · Manage SSH sessions and file transfers · Connect external coding agents through an API and skill

Interfaces: Web UI · CLI · SSH · Docker · HTTP API

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @Writbase

⭐ 11 stars · Growth: 9d +1 stars (+10.0%)

⚙️ Development Frameworks

An MCP-native task control plane for AI agent teams, with persistent tasks, scoped permissions, delegation, and provenance.

github

Capabilities: Maintain a persistent task registry for agents and human supervisors · Delegate tasks between agents with scoped permissions · Track task history and provenance

Interfaces: MCP server · CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @OpenBMB

⭐ 8,552 stars · Growth: 9d +3 stars (+0.0%)

🤖 AI Agents

XAgent is an open-source, experimental Large Language Model-driven autonomous agent designed to autonomously solve a wide range of tasks with features like autonomy, safety, extensibility, a GUI for easy interaction, and the ability to cooperate with humans

github | demo | docs | blog

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @xmacna

⭐ 1 stars · Growth: 9d +1 stars (+0.0%)

🤖 AI Agents

An installable AgentGet knowledge pack and Claude Code agent for designing AI sales workflows, including WhatsApp lead qualification, CRM hygiene, and supervised business automation.

github

Capabilities: Design lead qualification workflows · Define WhatsApp follow-up and handoff rules · Plan CRM data models and operating playbooks

Interfaces: AgentGet pack · Claude Code agent · Markdown instructions

Submitted by: @nikhonit submission accepted PR · Maintained by: not yet verified · Repository owner: @ZeroPointRepo

⭐ 1,004 stars · Growth: 9d +43 stars (+4.5%)

🤖 AI Agents

A collection of agent skills for retrieving YouTube transcripts, searching videos, browsing channels, and extracting playlists through TranscriptAPI.

github

Capabilities: YouTube transcripts · Video search · Channel browsing · Playlist extraction

Interfaces: Agent Skills

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @zapier

⭐ 260 stars · Growth: 9d +5 stars (+2.0%)

⚙️ Development Frameworks

The Zapier SDK is a TypeScript SDK and CLI for building agent tools and integrations against Zapier’s connection layer, giving coding agents programmatic access to Zapier’s app catalog and action library.

github | docs

Capabilities: Build agent tools and integrations · Discover connected apps and actions · Run actions through Zapier connections

Interfaces: TypeScript SDK · CLI

Submitted by: @Theaxiom (self-reported founder/team) submission accepted PR · Maintained by: not yet verified · Repository owner: @100monkeys-ai

⭐ 2 stars

🤖 AI Agents | 💻 Terminal-Friendly

Zaru is a pre-alpha open-source Rust terminal harness for LLM agents that runs the tools a model asks for under a permission model and records every step in a transcript.

github

Capabilities: Run a task against a Gemini, Ollama or OpenAI compatible model and execute the tools it requests, reading and editing files, running commands and fetching web pages, under a permission model · Write every step to a readable transcript · Iterate against project declared validators until they pass or a ceiling is reached

Interfaces: CLI

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @getzep

⭐ 4,948 stars · Growth: 9d +16 stars (+0.3%)

🧠 Long-Term Memory

Zep is a long-term memory service for AI assistants that enhances recall, understanding, and data extraction from chat histories to power personalized AI experiences

website | github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @ZeroPointRepo

⭐ 7 stars · Growth: 9d +1 stars (+16.7%)

🔧 Tool Calling (Function Calling)

MIT-0 agent skills for querying Zillow property data through Zillapi’s hosted API; the repository contains client skills, while the MCP server is hosted.

github

Submitted by: not yet recovered from repository history · Maintained by: not yet verified · Repository owner: @zooid-ai

⭐ 62 stars · Growth: 9d +4 stars (+6.9%)

🖥️ UI Development

Zooid is a self-hostable team-chat surface built on Matrix where people supervise AI agents running as teammates with approval cards, tool calls, and live plans.

github

Do you want to develop a custom agentic AI solution or looking launch your own AI agent?

Schedule a discovery call with me

Disclaimer: Slava Kurilyak is the Founder/CEO at Produvia

At Produvia, we partner with $7M+ in revenue brands to develop custom AI agents

Building an AI agent project? Prepare a project submission with the optional curation skill. New projects are welcome; there is no minimum star count.

To express your support, you can take the following actions:

1. Star the Repository: This helps increase its visibility.
2. Submit a Project: Founders can open a project submission issue; the optional skill can help prepare evidence and details.
3. Contribute Directly: Follow the contribution guide, then fork this repository and submit a pull request.

_{Disclaimer: issues and pull requests come from outside the maintainer team, so their content, including descriptions, attachments, code, and links, is treated as untrusted. Reviewers inspect the real destination of every link, leave submitted files and commands unrun, and check project claims against independent sources.}