Skip to content
Hot eventDeveloping

Qpilot releases AI agent that runs manual tests in Chrome

1 reports1 reporting sourcesUpdated 30 min ago

Start here

AI digest

Qpilot, released on GitHub by broxhq, is an AI agent that takes a pasted plain-text manual test case, executes each step in a real Chrome browser, and reports which step broke, showing expected versus actual results and a screenshot. It installs via npx qpilot and requires Node.js 20.12+, Chrome, and either an Anthropic API key or an OpenAI-compatible endpoint that supports tool calling. Each step ends as pass, fail or warn. A separate model call re-checks every pass and can only object, which downgrades the step to warn. Qpilot runs locally: only model calls leave the machine, so it can test staging and internal networks. Run history is kept in memory and limited to the last 50 runs, cleared on server restart, and the local API answers only Qpilot's own pages from localhost or 127.0.0.1.

Generated from reports · updated 9 min ago

Report timeline

Follow the reports to see the event from different sides.

November 18
  1. Hacker News · AI
    Qpilot launches an AI agent that runs plain-text manual test cases in Chrome

    Qpilot is a new AI agent that takes a pasted plain-text manual test case, executes each step in real Chrome, and reports which step broke with expected versus actual results and a screenshot. It installs via npx qpilot and needs Node.js 20.12+, Chrome, and an Anthropic API key or any OpenAI-compatible endpoint that supports tool calling. Each step ends as pass, fail or warn, and every pass is re-checked by a separate model call that can only object, turning the step into a warn.

Heat for this event

Not enough continuous observations yet to draw a trend.