Tell Claude, Cursor, or Copilot to test your app and write a guide — automatically. Or record manually, track bugs, and export visual docs. Free, privacy-first Chrome extension.
Sentinel replaces scattered tools with one lightweight extension. Use it manually or connect your AI assistant via MCP to fully automate testing and documentation.
Captures clicks, typing, scrolls, form submissions, and navigation automatically with intelligent noise reduction and smart selectors.
Replay recorded sessions at 0.5x to 4x speed. Step-by-step mode lets you pause after each action with element highlighting.
Add assertions to any step — visible, hidden, text-contains, text-equals, has-class, or exists. Get a clear pass/fail test report.
Turn recordings into polished, standalone HTML guides with embedded screenshots. Edit titles, reorder steps, and share anywhere.
Auto-capture console errors, unhandled exceptions, failed network requests, and CSP violations. Export styled issue reports.
Capture up to 5 minutes of video per clip as WebM. Download clips or attach them to your bug reports for clear reproduction.
Connect Claude, Cursor, or Copilot via the Model Context Protocol. Tell your AI to "test the login flow and write a guide" — it navigates, records, and exports automatically.
No setup, no accounts, no configuration files. Install, open the side panel, and start recording.
Click the record button and interact with your web app. Sentinel captures every action with smart selectors and screenshots.
Browse your recorded steps in the side panel. Each step shows what happened, where, and when — with a screenshot.
Add assertions, then play back the session. Sentinel simulates every action and evaluates your assertions, producing a test report.
Generate a standalone HTML guide, an issue report, or download a video clip. Share with your team — no logins needed.
Whether you're a solo developer or part of a QA team, Sentinel fits your workflow.
Record critical user flows, add assertions, and run regression playback before every release. Export test reports as evidence that paths work.
Walk through your app once and export a visual step-by-step guide. New team members get screenshots and descriptions instead of outdated wikis.
Enable error tracking, reproduce the bug, and export an issue report with screenshots, console errors, and network failures attached.
Record a walkthrough of completed features and export an HTML guide. Clients see exactly what was built — no staging environment needed.
Point your AI at a project folder and say "test the checkout flow and write a guide." Claude Code, Cursor, or Copilot navigate, record, assert, and export — hands-free via MCP.
Run a quick AI investigation on your dev server to surface JS errors and network failures before opening a PR. Get a screenshot + error report in seconds.
Choose your preferred installation method.
Read the full user guide to get started — from manual recording to AI-driven automation via MCP.
Sentinel is free and open source. If it saves you time, consider sponsoring the project ♥