AI Not Working? Identify the Failure Layer First

AI not working does not always mean the AI service is down. If ChatGPT, Cursor, Lovable, Replit, or another AI tool opens normally but your app keeps breaking, the issue may be inside your codebase, deployment, auth, database, or architecture — not the AI service itself.

The question is not only whether AI is down. The real question is whether your project should be fixed, rebuilt, handed off, or stopped before more AI-generated changes make the damage harder to recover.

Initial verdict

Do not assume the tool is broken.

medium risk

If the AI tool still opens but your generated app or code keeps failing, you are likely dealing with a build failure layer rather than a generic outage. Before prompting again, identify whether the damage is local, structural, deployment-related, or tied to auth, database, or permissions.

Tool issue vs build failure

If the tool opens but the project keeps breaking, the failure is often inside the build, not the AI service.

If the AI tool still responds but your app keeps getting worse after every prompt, stop treating it as a tool problem.

Tool outage

The service is down, rate limited, or failing to generate at all. Check status, account state, and whether the model can answer a simple prompt.

Account or browser issue

The tool opens inconsistently because of login, cookies, extensions, local browser state, or network filtering rather than your codebase.

Prompt or context issue

AI has partial information, weak task boundaries, or too much chat history, so each fix is local and unstable.

Project complexity issue

The request crossed from simple generation into state coordination, auth, payment, deployment, or long-lived maintenance.

Architecture or deployment failure

The code can compile or partially work, but the real failure sits in system boundaries, data flow, environment mismatch, or production setup.

AI-built app failed

Start here when the whole project keeps breaking after AI edits.

AI-generated code not working

Use this when AI answers, but the generated code does not run safely or reliably.

Build problems

Open diagnosis pages for code, deployment, auth, database, and data-layer failures.

AI tools

Compare Cursor, Lovable, Bolt, Replit AI, ChatGPT, and Claude Code failure types.

ChatGPT not working

Use this if you suspect the tool, browser, account, or network is failing.

Submit a stuck AI app

Request a structured decision before prompting again, hiring again, or launching.

Diagnosis

Need a fix-or-rebuild judgment?

Submit a stuck AI app for review before spending more time on the wrong fixes.

Second technical judgment: no secrets required.

Submit a stuck AI app