AI Agent
Sentinel
An open-source QA agent that reads codebases, derives business flows, and runs end-to-end tests across frontend and backend.
Sentinel
What is Sentinel?
Sentinel is an open-source AI agent that automatically reads your codebase, derives critical business flows, and tests them end-to-end by driving a browser and making backend assertions, catching integration bugs that UI-only checks miss.
Sentinel vs Similar AI Tools
| Pricing Model | Free | Free | Free, Free Trial | Free, Paid |
| Free Credits | ||||
| Key Features |
|
|
|
|
| Pros |
|
|
|
|
| Cons |
|
|
|
|
| Best For |
|
|
|
|
How to use Sentinel?
- 1Clone the repo from GitHub.
- 2Install dependencies with npm install.
- 3Copy config templates and set environment variables.
- 4Configure targets.json with your app's repo and credentials.
- 5Run bin/sentinel doctor to verify setup.
- 6Execute bin/sentinel run <app-name> qa to start a QA run.
Sentinel Key Features
- Reads repository structure without model calls (grep+find) to extract routes, services, and entities
- Derives prioritized end-to-end business flows from code analysis
- Runs each flow as an agent loop with Playwright browser driving and backend API assertions via token reuse
- Executes each flow multiple times (default 2) to handle non-determinism and unions findings
- Multimodal vision pass for design review (hierarchy, spacing, contrast, broken states)
- Supports wallet-gated apps via injected EIP-6963 provider (burner key, never broadcasts)
- Scheduled runs via launchd; caches flow plans per commit; open source under MIT
Sentinel Use Cases
- Automated QA for full-stack web applications
- Catching backend-frontend mismatches and state machine bugs
- Testing wallet-based dApps without risking real funds
- Continuous regression testing on every commit or scheduled interval
Sentinel Pricing & Free Credits
Sentinel currently operates on a Free model.
This tool is completely free to use
Sentinel Pros & Cons
Pros
- Open source and self-hosted, no vendor lock-in
- Automatically understands application logic from code, not just UI
- Validates both frontend and backend, finding integration bugs
- Supports wallet-based apps with safe injected wallet
- Provider-agnostic model configuration
Cons
- Requires manual setup of environment and dependencies
- Not a hosted SaaS; operator must manage infrastructure
- Agent exploration can vary between runs, requiring multiple attempts
- Depth limited by configurable budgets (steps, attempts); may miss edge cases on large apps
What is Sentinel best for?
- Development teams needing automated end-to-end QA without writing test scripts
- Projects with complex business flows where UI-only testing misses backend state bugs
- Teams wanting to catch regressions early in CI without manual test plans