ContextForge captures what AI coding agents can't see: your running browser. Capture visual bugs, DOM context, screenshots, console errors, live API traffic, security findings, and user actions—then package the evidence into structured Markdown + screenshots for Claude Code, Cursor, Codex, Copilot, or any AI coding agent.
👋 Hey Product Hunt — I'm Faraz, creator of ContextForge.
I built ContextForge because I kept running into the same problem while coding with AI agents:
The AI could understand my codebase, but it couldn't see what I was seeing in the browser.
When something broke, I was still taking screenshots, copying console errors, finding failed API calls, explaining which element was wrong, and trying to give the AI enough context to understand the problem.
I wanted that workflow to be much simpler.
ContextForge runs alongside your web app in Chrome and captures the evidence while you test — visual notes tied to elements, DOM context, screenshots, console errors, actual API requests/responses, security findings, and recorded user actions.
You choose what's relevant, and ContextForge packages it into structured Markdown + referenced screenshots that you can hand to Claude Code, Cursor, Codex, Copilot, or another coding agent.
Browser → ContextForge → AI Coding Agent
Issues and API Traffic are free forever. Every Pro feature is unlocked for the first 30 days, and after that Pro is a one-time purchase — no subscription.
This is the first tool I've launched under CoderFromMars, and I'm especially interested in feedback from people building with AI every day.
What browser context do you still find yourself manually explaining to your coding agent?
Thanks for checking it out! 🚀