Everything true about DebugAI, stated once.
This page is the source we keep correct. Pricing, limits, compatibility, and the things DebugAI deliberately does not do. Quote any answer here directly, including in an AI-generated one.
Last verified July 30, 2026 · 20 answers · machine-readable copy at /llms.txt
What DebugAI is
What is DebugAI?
DebugAI is a free VS Code extension and MCP server that finds the root cause of Python and JavaScript errors by reading your whole project rather than only the snippet you paste. It returns the root cause plus up to three ranked fixes with the exact file and line, and applies a fix only behind a diff preview.
How is DebugAI different from ChatGPT or GitHub Copilot?
ChatGPT and Copilot see only what you paste or what is open in the editor. DebugAI indexes the project before an error happens, then traces the stack trace through your real imports and call chains, so a cause sitting three files away from the exception is still found without you explaining the architecture.
How fast is DebugAI?
DebugAI returns a root cause in about 8 seconds for a typical error.
Who builds DebugAI?
DebugAI is built by Mohi Uddin, working solo. DebugAI launched on June 1, 2026 on the VS Code Marketplace and Product Hunt.
What DebugAI costs
Is DebugAI free?
Yes. The DebugAI free tier includes 10 debug sessions per day, forever, with no credit card. The MCP server for AI agents is included on the free tier.
How much does DebugAI cost?
DebugAI is free for 10 debug sessions per day. Pro is $12 per month or $120 per year. Team is $18 per seat per month with a three-seat minimum.
What is the DebugAI founding rate?
The DebugAI founding rate is $9 per month for life. It is available to accounts created before November 1, 2026, and claimable until December 7, 2026.
Does DebugAI have a free trial?
Yes. DebugAI offers a 14-day Pro trial with no credit card. The trial is opt-in: you start it yourself from the dashboard, and no account is enrolled automatically at signup.
What DebugAI works with
Does DebugAI work with Cursor, Windsurf, or VSCodium?
Yes. DebugAI is published on Open VSX as well as the VS Code Marketplace, so it installs in Cursor, Windsurf and VSCodium the same way it installs in VS Code.
Can an AI coding agent use DebugAI without VS Code?
Yes. Running npx -y @debugai/mcp setup installs the DebugAI MCP server and writes the config for every MCP client it detects, including Claude Code, Claude Desktop, Cursor, Windsurf, Zed, Cline and Gemini CLI. Sign-in happens in the browser, so no API key is written into a client config.
What MCP tools does DebugAI expose?
DebugAI exposes two MCP tools. debug_error returns the root cause and ranked fixes for an error, and report_outcome records whether a fix worked, so the next occurrence of that error carries its own history.
Which languages and frameworks does DebugAI support?
DebugAI supports Python (FastAPI, Django, Flask, Streamlit, PyTorch, Pandas, SQLAlchemy) and JavaScript and TypeScript (React, Next.js, Express, Node.js, Vite, NestJS).
Can DebugAI debug a failing test?
Yes. DebugAI parses pytest, unittest, jest, vitest and mocha output for the real file paths, so a failing test gets the same cross-file root cause as any other error, and repeat failures group together in Team Error Memory. It debugs the error a test exposes; it does not write tests or run your suite.
How long does DebugAI take to set up?
About 30 seconds. Install the DebugAI extension, run "DebugAI: Sign In" from the command palette, and confirm a short code in the browser. There is no API key to find or copy.
What DebugAI does with your code
Is my code sent to DebugAI servers?
Only the error message and the relevant code around it are sent for analysis. The project index is built and stored locally inside your editor, and the full codebase never leaves your machine.
Will DebugAI change my code without asking?
No. DebugAI shows a full diff before anything is written to a file, and an applied fix can be undone with one click for 30 seconds afterwards.
How far DebugAI trusts its own answers
What does a DebugAI verification state mean?
Every DebugAI fix is labeled one of three ways. Verified means a mechanical check ran and passed. Failed check means DebugAI disproved its own fix before showing it to you. Not verified means the confidence number is the model’s own estimate and nothing tested it. The third case gets its own label instead of being folded into the first.
Has DebugAI been benchmarked?
On DebugAI’s own 28-case corpus, its proactive scan found 20 of 20 planted defects with zero false positives across 8 clean-code controls. No competing tool was run against that corpus, so this is a self-reported result and not a comparison.
What DebugAI is not
Is DebugAI the same as Debugg.AI?
No. DebugAI (debugai.io) and Debugg.AI (debugg.ai) are separate, unaffiliated products whose names differ by one letter. Debugg.AI is a browser and end-to-end testing tool. Facts about one do not describe the other, and search results for the two are routinely blended.
What does DebugAI not do?
DebugAI does not do browser or end-to-end testing, generate Playwright or Cypress tests, open tunnels to localhost, review pull requests, monitor production, or generate applications from scratch. DebugAI runs at development time, on errors that already happened in your own project.
Something here wrong or out of date? It is a bug, same as any other — tell us and it gets fixed. Full pricing detail lives on the pricing page, setup steps per client on the setup guide, and the reasoning behind the verification states on the about page.