Open-Capture/opencapture

The open-source alternative to GoFullPage and FireShot — no access to your other tabs, nothing uploaded. opencapture.app

A privacy-first, 100% local full-page screenshot and annotation extension for Chrome, Edge and Firefox.

Capture a full page, the visible viewport, or a hand-picked region — then crop, arrow, rectangle, blur, or watermark it, and export as PNG or PDF. Every pixel operation happens on your machine, in a Rust/WebAssembly core — nothing is uploaded anywhere, no account is needed, and the capture pipeline never has standing access to any page you haven't explicitly asked it to capture. Screenshots, docs and install links are at opencapture.app, the OpenCapture website.

Most "free" screenshot tools fund themselves one of two ways: your images pass through their servers (openly, for sync/sharing — or, historically, less openly), or the free tier is a funnel toward a subscription. Neither is inherently malicious, but it means your screenshots — which may contain account numbers, internal dashboards, private conversations, anything that happened to be on your screen — leave your machine by design.

Your screenshots have no server to go to. The manifest requests no host_permissions and no static content script; capture code is injected on demand, only into the tab you click the icon on, and every image stays in the browser's own storage until you explicitly download, copy, or annotate it.

The one request that can leave this extension goes to our own account server, and only while you are signed in — see Permissions for which hosts those are and what asks for them. Signing in is optional and nothing about capturing, editing or exporting depends on it.

• Three capture modes — full page (auto-scroll + stitch, pixel-exact across device pixel ratios), visible viewport only, or a hand-dragged region.
• Annotate before you export — crop, arrow, rectangle, text, and a genuinely irreversible mosaic blur (not a cosmetic CSS filter — the source pixels are gone), each with full undo.
• Watermark tool — tile your own text and/or logo across the top, bottom, both, or the entire page, at 0° or a 45° slant. This is a tool you apply to your own exports for your own branding — OpenCapture never stamps anything onto your screenshots itself. The one paid part of OpenCapture: a one-time, 1000-credit Supporter unlock. Every other feature — capture, crop, arrow, rectangle, text, blur, PNG/PDF export — is free with no account at all.
• Export PNG or PDF — PDF pages are sized in points to match CSS pixels, so they print at true 1:1 physical size.
• Copy to clipboard, configurable save folder/filename, and a capture that survives the background service worker being evicted mid-flow.
• Chrome, Edge and Firefox, same source tree, two build targets.

Every figure is checked against that product's own site or store listing. Last reviewed October 2026 — tell us if something is out of date.

Full comparison, including where each one is the better pick →

Also on opencapture.app.

Official listings. OpenCapture is published in exactly three places:

• Chrome: Chrome Web Store
• Edge: Microsoft Edge Add-ons
• Firefox: Firefox Add-ons

Other screenshot extensions use similar names. If an add-on is not at one of these three addresses, it is not this one, and we cannot help with it.

Getting help. For a bug or a feature request, open an issue and say which browser and version you use. For anything else, email hello@opencapture.app. The FAQ answers the common questions.

If OpenCapture saved you time, a star helps other people find it.

Published and feature-complete: full-page, visible-area and selected-region capture, crop/annotate/watermark, and PNG/PDF export all work today, on all three stores.

To run a local build instead — for development, or for Android — see docs/LOCAL_TESTING.md, docs/FIREFOX_TESTING.md and docs/MOBILE_TESTING.md.

Four permissions are requested at install, and they are the whole set:

• activeTab — read the page you are capturing, and only the tab whose toolbar icon you clicked, only after you click it.
• scripting — inject the capture script into that tab on demand. There is no static content script, so nothing runs on any page until you ask.
• downloads — write the PNG or PDF you asked to save.
• storage — remember your preferences and the local capture history.

Three more are declared in optional_host_permissions, which means the browser does not grant them at install and never asks unless you take the action that needs one:

• auth.opencapture.app, gateway.opencapture.app — our account server. Requested only if you choose to sign in, which is needed for the one paid tool and nothing else. Never signing in means never granting these.
• openpdfedit.com — requested only when you tick "Open the PDF in OpenPdfEdit" while exporting, and it exists so the finished PDF can be handed straight to that editor instead of going out through your downloads folder. Leave it unticked and the permission is never asked for.

# prerequisites: Rust (stable) + wasm32-unknown-unknown target, Node 22.x,
# wasm-bindgen-cli pinned to exactly 0.2.100 (must match Cargo.toml)
rustup target add wasm32-unknown-unknown
cargo install wasm-bindgen-cli --version 0.2.100 --locked

cd apps/extension
npm install
npm run build            # -> dist/          (load unpacked in Chrome)
npm run build:firefox    # -> dist-firefox/   (load temporarily in Firefox)

Then in Chrome: chrome://extensions → enable Developer mode → Load unpacked → select dist/. In Firefox: about:debugging#/runtime/this-firefox → Load Temporary Add-on → select any file inside dist-firefox/.

The tiled watermark is not part of this repository. It is a separate private module, it is not covered by this repository's licence, and a clone of this repository builds without it: the build substitutes a stub and says so, every other tool works normally, and only the watermark tool does nothing. Nothing else here depends on it.

Full walkthroughs, including the shared-mount build gotcha and what to manually click through, are in docs/LOCAL_TESTING.md, docs/FIREFOX_TESTING.md, docs/MOBILE_TESTING.md, and docs/TESTING_GUIDE.md.

Rust does every pixel operation; TypeScript is the browser-API shell around it.

• crates/shot-core — scroll-target planning, gap/overlap-free slice stitching, PNG decode/stitch/encode, multi-page PDF export, oversized- canvas splitting. Compiles natively and to wasm32-unknown-unknown via wasm-bindgen.
• crates/shot-qa — native CLI for PNG/PDF structural inspection, hashing, and pixel sampling/diffing.
• apps/extension — the MV3 shell (TypeScript, Vite): background service worker orchestrates capture, an on-demand content script handles DOM/lazy-load/sticky-element prep, and the popup/editor pages are plain canvas-based TypeScript.

The design rationale for each part lives in comments beside the code, and the build and test steps are in docs/LOCAL_TESTING.md.

OpenCapture never requires an account for any capture, annotation, or export feature — everything above works fully offline. An optional sign-in exists in the popup, and one tool — the tiled watermark — is unlocked through it. Declining it changes nothing else: every capture mode, every annotation tool and both export formats work without an account, offline.

AGPL-3.0-or-later. If you build on this code and offer it as a network service, the AGPL requires you to make your source available to that service's users too.