Self-hosted QA platform for IDE agents: generate, run, and publish E2E tests with video evidence from Claude Code or Cursor. Works with or without an LLM.

Hi Product Hunt! 👋 I built Suitest because I was frustrated with the current state of AI-assisted testing. While AI coding agents are incredible, relying on pure AI to write and run end-to-end tests often leads to flaky results and constant hallucinations. On top of that, existing agentic testing platforms (like TestSprite) use expensive, credit-based pricing. I wanted a reliable, free alternative. At its core, Suitest is a web platform that automates the entire E2E testing lifecycle: analyze → generate → start → wait → run → report. Instead of guessing, it uses deterministic generators via OpenAPI, web crawlers, and browser recorders. This guarantees your tests are grounded in reality, not AI hallucinations. But we didn't stop at the web UI. We also built a native MCP (Model Context Protocol) server. This means you can automate your tests directly via our web platform, OR you can plug our MCP server straight into your IDE agents (like Claude or Cursor). This allows your AI assistant to seamlessly orchestrate a rock-solid, deterministic testing flow right from your editor without the expensive pay-per-action fees. I’d love to hear your feedback on the architecture, how you currently handle E2E testing, or any features you'd like to see next!

Spun it up locally and pointed Claude Code at my app, the video evidence on each run is such a nice touch for catching flaky UI bugs. Works without the LLM too which is great for CI.

Ran it with claude code and the video evidence after each test run is genuinely useful for catching flaky behavior. Like that it works without an LLM too, makes it easy to slot into existing CI.

Spun it up locally and pointed Claude Code at my app, the video evidence on each run is such a nice touch for catching flaky UI bugs. Works without the LLM too which is great for CI.

Ran it with claude code and the video evidence after each test run is genuinely useful for catching flaky behavior. Like that it works without an LLM too, makes it easy to slot into existing CI.

Suitest

Free E2E AI testing & MCP server. A TestSprite alternative.

Free E2E AI testing & MCP server. A TestSprite alternative.