Bug-Zap is MCP-connected testing for your app. Your AI understands the code and proposes the tests; Bug-Zap runs them in a real browser, verifies every assertion, and keeps replayable screenshots and traces.
Bring your own AI model · unlimited local test runs on every plan
The AI understands your code. Bug-Zap stores, runs, and verifies.
Connect Claude Code, Cursor, or any MCP client. Your AI reads the code, proposes tests, and submits them straight into your project.
Playwright drives an actual browser — local by default, cloud when you need it. Combine UI steps with API assertions in one test.
Every run keeps screenshots, traces, video, console/network errors, and a step-by-step timeline you can replay and compare.
Saved tests run as deterministic recipes — no active AI session required. Re-run after every change and catch what broke.
Staging-first. Production is read-only unless you approve a scoped action. Secrets are redacted from logs, traces, and screenshots.
Organize tests into draggable groups and cases. See last status at a glance; drill into full history and evidence per case.
Add the Bug-Zap MCP server to Claude Code or Cursor with a copy-paste config.
Your AI inspects the codebase, proposes project-specific tests, and submits them over MCP.
Execute locally in an isolated Playwright browser or in the cloud — single case, group, or whole suite.
Get truthful pass / fail / blocked status with screenshots, traces, and replayable step timelines.
Unlimited local test runs on every tier. No surprise cloud charges.
For solo developers
For growing products
For teams
Let your AI write the tests and let Bug-Zap prove they pass — with evidence you can replay.
Get started