Last updated 08-30-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Debugg AI
Debugg AI runs managed browser tests on your web app and posts pass or fail results where your team already works, mainly GitHub pull request comments with session recordings. Connect a repo through the GitHub App and the platform handles cloning, builds, secure tunneling, and real browser sessions without local Playwright or Selenium setup on your machine.
Where most E2E stacks make you author and maintain test scripts, Debugg AI explores pages and flows like a real user, then runs checks on signup, checkout, login, and other revenue paths on every change. You also get on-demand runs against localhost through auto-created tunnels, plus an MCP server so Claude, Cursor, or Codex can call check_app_in_browser against the build on your desk.
It fits solo developers, open source maintainers, and QA-minded teams who want UI regressions caught before merge. PR Copilot and VS Code or Cursor extensions bring the same testing into your editor, while AI Test Suites turn observed usage into framework-ready tests for Jest, Pytest, and similar runners.
GitHub App connects in about two minutes and posts E2E results as inline PR comments with video recordings
Free plan covers public repos with 100 tests per month; Pro is $20 per month for 1,000 tests on private repos
Managed remote browsers reach localhost through an auto-created secure tunnel with no manual ngrok setup
MCP server ships 8 tools and 21 actions, including check_app_in_browser for Claude Desktop, Claude Code, and Cursor
AI Test Suites generate Jest, Mocha, Pytest, and JUnit tests from observed app usage without hand-written scripts
On-demand runs spin up a real browser in seconds and replay failures step by step with full session recordings
PR Copilot extension for VS Code and Cursor reviews commits and runs generated browser tests inside the editor
Posts browser test results directly in GitHub pull request comments with video recordings.
No local browser automation setup required thanks to managed remote browsers and auto-created localhost tunnels.
Free tier covers public repositories with 100 tests per month before any paid plan.
MCP server and VS Code or Cursor extensions let coding agents verify UI flows without leaving the editor.
Private repositories and higher test volumes require a paid plan starting at $20 per month.
Primary workflow is GitHub-centric, so teams on other Git hosts need to evaluate fit.
Enterprise features such as SSO/SAML require contacting sales on a custom plan.
Is Debugg AI free to use?
Yes. Debugg AI offers a Free plan for public repositories with 100 tests per month, PR comments, and community support. Pro costs $20 per month for private repos and 1,000 tests, and Grow is $40 per month for 5,000 tests.
How does Debugg AI know what to test?
Debugg AI explores your app like a real user, covering pages, buttons, and flows customers rely on. It then runs targeted browser checks on signup, login, checkout, and other critical journeys whenever a change lands, instead of replaying one fixed script every time.
Do I need to configure Playwright locally?
No. Debugg AI provides fully managed remote browsers and handles repository cloning, builds, tunneling, and test execution. The marketing site describes setup as taking about two minutes through the GitHub App with no code changes required.
Can Debugg AI test my local dev server?
Yes. Debugg AI can point at a URL such as http://localhost:3000 and opens a secure tunnel so managed browsers reach your running app. The MCP check_app_in_browser tool uses the same tunneling so agents never see the tunnel URL in tool output.
Does Debugg AI work with MCP clients like Claude and Cursor?
Yes. Install the Debugg AI MCP server with npx @debugg-ai/debugg-ai-mcp and an API key from app.debugg.ai. It exposes 8 tools and 21 actions for Claude Desktop, Claude Code, Cursor, LangChain, and other MCP clients.
Is there a VS Code or Cursor extension for Debugg AI?
Yes. Debugg AI offers PR Copilot extensions for VS Code and Cursor that review commits and run generated browser tests inside the editor. The same platform also posts E2E results to GitHub pull requests when you use the GitHub App workflow.