Project Details
theORQL
º2026
Project Overview
A coding assistant that lives in your editor and browser. It builds projects, writes files, resolves problems, creates layouts, and navigates repositories. Its built-in browser debugger launches a project URL and injects the chat directly into the page, letting the AI read console logs, inspect the DOM, execute JavaScript, take screenshots, and analyze images to fix frontend errors without switching windows. The project was built by the theORQL team with our founder participating as a full stack developer during its early stages, contributing to the foundation and leading core feature development alongside their team.
Challenge
Integrating artificial intelligence into something as complex as frontend error debugging. We noticed developers wasted time copying and pasting logs into AI tools, so we set out to integrate theORQL — which lives inside VSCode — with the browser to capture console errors. We achieved this, and even added the chat UI directly in the browser. Then we wanted to expand our debugging capabilities, so the challenge became automating the frontend debugging process using AI, with the goal of saving time.
Results
We created an assistant that, unlike others, has context and vision of the frontend. It can read logs, execute JavaScript, take screenshots, read the DOM, read requests, and is equipped with filesystem skills — giving us the ability to use it as a programming tool, not just for debugging. We also automated the error debugging process by converting it into 3 phases: first, the AI replicates the error using browser tools like clicks, fills, and execute to see the console output (this is crucial because it's how the AI verifies if the issue was resolved); after replicating, it analyzes the code to find the problem, uses the replication information to understand it, and applies the fix; finally, it replicates the error again to verify the result, and if not resolved, enters a loop with the context of the previous fix.