astrid-runtime/astrid

Software should compose without inheriting each other's authority.

Astrid is an operating-system project for software built from WebAssembly capsules. Today it runs as a portable user-space runtime on macOS and Linux; the direction is a standalone operating system. This release is the hosted runtime, not a bootable OS image.

Run components in a sandbox, connect them through typed interfaces, and give each principal its own capabilities and durable state. Install, upgrade, and remove capsules while the runtime is running.

Use it to build agent systems, tools, services, or your own distribution. Astrid does not choose a product, model provider, agent loop, or user interface for you.

Agents make the problem urgent: a model can choose what to do, but that should not make the model the authority that permits it. Astrid separates those jobs. The same foundation serves software without an LLM—components can be replaced and composed without putting application policy into the kernel.

• Compose behavior: capsules communicate through an event bus and versioned interfaces; the kernel routes requests rather than owning application logic.
• Make authority explicit: filesystem, network, process, and IPC access are checked at runtime boundaries, not entrusted to prompts.
• Keep state portable: content-addressed storage, deduplication, and recovery back principal-owned data. Mount it through macOS FSKit or Linux FUSE.

Get started · Write a capsule · Read the Book · Release notes

Install Astrid using one of the options below, then start an uncomposed runtime—no model account or distribution required:

astrid --version
astrid start
astrid status
astrid capsule list
astrid stop

status reports the running daemon; a fresh uncomposed runtime has no product capsules. stop shuts it down and retires its working state into astrid.volume. Starting again restores the working projection.

Next, build your first capsule, or choose a distribution you trust with astrid init --distro <source>. A distribution supplies the capsule composition and configuration; Astrid does not select one implicitly. See initial setup for repository, local-manifest, and signed-bundle inputs.

Start with the Book for the architecture or the Contributor Handbook to contribute.

Agent frameworks put trust in the prompt. Astrid puts it in the runtime. An agent is untrusted code executing on your machine with access to your files, your network, and your credentials. Telling it to behave is not a security boundary. An OS-grade boundary is.

• Cryptographic capability model. Every file path, network host, and tool is a signed ed25519 grant scoped to a resource pattern, principal-bound, expiry-checked, and globally revocable. No grant, no access.
• WASM sandbox with no ambient authority. Capsules run in Wasmtime with no syscalls, no file descriptors, and no host memory. Every external effect is a capability-checked host call over a WIT-typed ABI.
• The kernel is dumb. It instantiates an event bus, loads capsules, and routes IPC bytes under a capability ACL. It has no LLM handles, no conversation state, and no tool registry. All intelligence lives in capsules, so a capsule bug cannot corrupt shared kernel state.
• Per-principal everything. Each identity gets isolated capsule access, KV data, secrets, home directory, quotas, and audit chain. One principal can never read another's namespace, and it fails closed if the caller cannot be resolved.
• Signed, hash-linked audit chain. Each entry seals the hash of the one before it and is signed. Break the chain and the tampering shows.
• Live capsule lifecycle. Install, upgrade, and remove capsules on a running daemon. No restart.

Frontends (the CLI, the HTTP gateway, Discord, and so on) are uplinks: protocol clients that connect to the daemon over a Unix domain socket and speak in IPC events. There is no Frontend trait. An uplink publishes events and receives responses like any other bus participant.

flowchart TB
    CLI[CLI] -->|IPC events over Unix socket| Kernel
    HTTP[HTTP gateway] -->|IPC events| Kernel
    Discord[Discord] -->|IPC events| Kernel
    Uplinks[Other uplinks] -->|IPC events| Kernel

    subgraph Runtime[Astrid]
        Kernel[Kernel: astrid-daemon<br/>dumb event router<br/>event bus - capability ACL - audit chain - Wasmtime sandbox]
        Capsules[WASM Component capsules: wasm32-unknown-unknown<br/>providers - orchestrators - tools - fs - http - sessions - registry - identity]
    end

    Kernel -->|capability-checked host calls<br/>astrid:* WIT ABI| Capsules

Capsules communicate exclusively through the bus. Each declares what it needs and what it provides in a Capsule.toml manifest with typed [imports]/[exports] tables; the kernel resolves the dependency graph by topological sort and boots capsules in order. Tools are an IPC convention, not a kernel concept: a tool capsule intercepts tool.v1.execute.<name>, and the kernel never sees a tool schema.

The host ABI is the WebAssembly component model with versioned astrid:* WIT packages: fs, io, kv, ipc, net, http, sys, process, approval, identity, elicit, and uplink. Guests import only what their manifest allows, and every call is capability-gated at the boundary.

Astrid's security is decomposed. There is no single gate every action funnels through. A capsule has no ambient authority, and authorization is enforced by independent, per-area mechanisms, each fail-closed and each enforced where the effect actually happens.

flowchart TB
    Action[A capsule action] --> Sandbox[WASM sandbox<br/>no syscalls, file descriptors, or host memory<br/>external resources are capability-checked host calls]
    Action --> Manifest[Manifest gate<br/>declared file, network, and process allow-list<br/>empty is deny-all; path traversal and SSRF defenses]
    Action --> IPC[IPC ACL<br/>declared publish and subscribe topics only<br/>per-principal routing]
    Action --> Capability[Capability token<br/>ed25519, principal-bound, scoped, expiring, revocable<br/>per-device tokens can be subset-scoped]
    Action --> Approval[Approval gate<br/>once - session - always - deny<br/>allow always mints a token; local egress elicits consent]
    Action --> OS[OS sandbox<br/>bwrap on Linux - seatbelt on macOS for native subprocesses]
    Action --> Audit[Audit chain<br/>signed, hash-linked JSONL decisions and calls<br/>split per principal and independently verifiable]

These mechanisms are real and independently tested. There is no unified interceptor orchestrating them. The five-layer gate chapter of The Astrid Book walks each layer against the source.

Release archives:

Download the archive for your machine from GitHub Releases. Keep its companion binaries and platform files together when extracting it; add the extracted directory to your PATH.

The current stable release, v2026.10.0, publishes these targets:

Windows is tested in CI but is not included in the release archives. For older runtime-layout migration and platform limitations, read the 2026.9.0 upgrade notes; for changes in this release, read the 2026.10.0 changelog. Versions follow year.month.patch.

Existing self-managed installations can run astrid update --check to inspect their signed channel, then astrid update to apply an available update. Homebrew and Cargo installations are updated by their respective package managers.

Homebrew (macOS and Linux):

brew tap astrid-runtime/tap
brew install astrid

From crates.io (requires Rust 1.95+):

cargo install astrid

From source:

git clone https://github.com/astrid-runtime/astrid
cd astrid && cargo build --release   # binary at ./target/release/astrid

The core tools are listed below. Platform release archives additionally include their filesystem provider and, on macOS, the signed AstridFS app and management scripts. A Cargo install is not a substitute for that signed macOS bundle.

astrid init --distro <source> fetches a distro (a curated capsule bundle), presents any selection groups declared by that distro, and prompts for its required configuration. Secrets are stored per principal in the secret store, never passed on the command line. init writes a Distro.lock pinning every capsule by BLAKE3 hash, so the same explicit distro input reproduces the same fleet.

astrid init --distro @yourorg/your-distro          # repository distro
astrid init --distro ./Distro.toml                  # local manifest
astrid init --distro ./bundle.shuttle --offline     # signed bundle, no network
astrid init --distro @yourorg/your-distro --yes     # non-interactive defaults

When a distro includes an LLM provider, onboarding can discover that provider's live model list from its /v1/models endpoint. If the distro also includes an agent loop, confirm that it is ready before starting a session:

astrid doctor    # daemon up? capsules ready? an LLM available?
astrid chat      # interactive session; the daemon auto-starts on first use
astrid models    # list the current provider's models (a registry-capsule verb)

astrid -p "summarize the git log"                    # single prompt, prints and exits
git diff HEAD~1 | astrid -p "write a commit message" # stdin is appended to the prompt
astrid -p "fix all failing tests" --yes              # auto-approve tool requests
astrid -p "continue" --session "$SID"                # resume by id or name

astrid start     # persistent daemon (survives terminal close)
astrid status    # PID, uptime, connected clients, loaded capsules
astrid ps        # loaded capsules and their lifecycle state
astrid stop      # graceful shutdown
astrid update    # authenticate, verify, and install the latest release

An explicit astrid start is persistent. Automatically started MCP gateways and ephemeral daemons instead follow client-connection lifetime: quiet connected clients keep them alive, and they retire after the final client disconnects.

Once the platform filesystem frontend is installed and enabled, a principal can mount its view through the CLI:

astrid start
mkdir -p /tmp/astrid-view
astrid --principal default storage mount --as default /tmp/astrid-view
astrid --principal default storage status /tmp/astrid-view
astrid --principal default storage sync /tmp/astrid-view
astrid --principal default storage unmount /tmp/astrid-view
astrid stop

Use an empty mountpoint. Principal views and the administrative runtime view have different authority; mounting does not grant access to other principals. The files live in Astrid's storage, backed by the host filesystem—not in a new disk partition. Release binaries remain outside the runtime's mutable volume.

macOS archives include macos/manage-macos-fskit.sh for app installation and enablement; follow its macOS permission guidance. Linux uses its FUSE frontend and host FUSE setup, without Apple's signing or extension-approval steps.

Each principal (agent identity) is a fully isolated tenant: its own capsule access, KV namespace, secrets, home directory, quotas, and audit chain. New principals inherit nothing by default.

astrid agent create ci-bot                     # clean-slate, least-privilege agent
astrid agent create staging --clone production # full profile + state replica
astrid agent modify ci-bot \
  --add-capsule astrid-capsule-fs              # grant access to a capsule's tools
astrid caps show ci-bot                        # inspect capability grants
astrid quota set -a ci-bot --memory 128MB      # per-principal resource limits
astrid pair-device issue --scope use-only      # scope a device token to a subset

Capsule access is enforced kernel-side at dispatch. A principal can only invoke capsules explicitly granted to it, and two principals installing the same capsule bytes share one content-addressed on-disk artifact while getting separate in-memory runtime instances.

A capsule is a WASM process described by a manifest. The scaffold generates a first-try-compiling project targeting wasm32-unknown-unknown, plus an AUTHORING.md guide.

astrid capsule new my-capsule    # scaffold Capsule.toml, Cargo.toml, src/lib.rs, .cargo/config.toml
cd my-capsule
astrid capsule build             # compile and package
astrid capsule install .         # hot-loaded into the running daemon, no restart

Capsule authors depend on astrid-sdk, which mirrors the std module layout (fs, net, process, env, time, log) and adds Astrid modules (ipc, kv, http, hooks, uplink, identity, approval). The #[capsule] proc macro generates the WASM ABI boilerplate: exports, serialization, and dispatch for tools, commands, hooks, and lifecycle entry points.

use astrid_sdk::prelude::*;

#[derive(Default)]
struct Weather;

#[capsule]
impl Weather {
    #[astrid::tool]
    fn forecast(&self, args: ForecastArgs) -> Result<Forecast, SysError> {
        let key = env::var("WEATHER_API_KEY")?;
        let body = http::get(&format!("https://api.example.com/wx?q={}", args.city))?;
        Ok(serde_json::from_slice(&body)?)
    }
}

Building capsules is a first-class workflow in Astrid. The Forge (astrid capsule new plus authoring tools and runtime-served guidance) is the on-ramp; the direction is an agent that writes, builds, and installs its own capsules within the capability sandbox.

Reproducible builds and deterministic tests are consumers of Muninn's verified computation memory: complete source and toolchain closures become derivation inputs, so an unchanged fleet build is a verified lookup while install, signing, and publication remain fresh authorized effects.

astrid capsule install @org/capsule-name   # install and hot-load
astrid capsule update my-capsule           # hot-swap the running instance
astrid capsule remove my-capsule           # live-unload, no restart
astrid capsule list --verbose              # installed capsules with capability metadata
astrid capsule tree                        # imports/exports dependency graph

• Native-tool readiness: Linux setup now checks ordinary-user Bubblewrap and FUSE prerequisites and explains repairs before approved tools fail. Multiple Unix users can start their own runtimes on a shared host.
• Workspace reliability: macOS Node can resolve authorized script files, and populated-workspace renames preserve pending content during the atomic transaction.
• Sandbox and dependencies: native process grants no longer expose runtime backing media; Rustls and Wasmtime include upstream security fixes.

See the full 2026.10.1 changelog for the exact fixes and operational limits.

• Accountable user budgets: operators can configure an optional shared WASM fuel-rate budget per accountable user in [resources], applied on restart. Principal-level astrid quota set --cpu-rate remains separate.
• Capsule update discovery: astrid capsule update --check --json reports available releases without installing them or granting new authority.
• Audit evidence: signed audit-head snapshots and paged exports make chain history inspectable; opt-in v2 entries add deterministic CBOR and committed fields while retaining existing v1 history.
• Release and recovery work: numbered dev-channel release candidates, corrected update notices, and repairs to storage migration, native mounts, nested capsule failures, and concurrent runtime work.

See the full 2026.10.0 changelog for the exact changes and migration limits.

• The Astrid Book is the canonical reference: the kernel, the capsule model, the host ABI, the bus, and the security model, grounded in the source with file and line anchors. Start with Getting Started to go from nothing to a working agent in a few minutes.
• Operator guides live in docs/: the unified config schema, LLM model selection, distro signing, the gateway deployment runbook, and generating a gateway client. Architecture programs include conservation of computation, Muninn, Huginn, and the Refinery.

cargo build --workspace
cargo test --workspace -- --quiet
cargo clippy --workspace --all-features -- -D warnings
cargo fmt --all -- --check

All crates enforce #![deny(unsafe_code)] except astrid-sys and astrid-sdk, where WASM FFI requires it. Clippy runs at pedantic level and integer-overflow arithmetic is a lint error. Release binaries for macOS and GNU/musl Linux (x86_64 and aarch64) are built on tag push and signed with keyless Sigstore. Self-managed updates authenticate the exact archive and its pinned Astrid release-workflow identity before independently checking the BLAKE3 manifest and extracting any bytes. Homebrew and Cargo remain responsible for updates they install; signed SHA-256 manifests remain available for their compatibility requirements. GitHub build-provenance attestations are published as additional evidence and are not substituted for the updater's release-archive signature. See the self-update security model.

Contributions are welcome. Astrid uses a tiered contributor system that protects security-critical code while keeping the door open to new contributors. Every pull request must be linked to a GitHub issue. See CONTRIBUTING.md for the issue-first workflow and tier descriptions.

Changes to any contract surface (the host ABI, the IPC protocol, the capability model, the manifest schema, or the SDK public API) go through the RFC process in the RFCs repository before implementation.

Dual-licensed under MIT and Apache 2.0, at your option.

Copyright (c) 2025-2026 Joshua J. Bouw and Unicity Labs.