post_debug_code
LLM CODE DEBUGGING — POST {code, error} and get a diagnosis: what is wrong, the root cause, and a concrete fix with corrected code. Paste the failing snippet plus the error message or stack trace; any language, up to 20,000 chars combined. Optional {language} and {context} ('happens only on the second call'). Fast cheap LLM under the hood. Want deterministic no-AI lint instead? POST /api/lint/:language ($0.002). ($0.01 per call, paid via x402)
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| code | Yes | The failing code snippet (required) | |
| error | No | Error message, stack trace, or a description of the wrong behavior | |
| context | No | Optional extra context: what you expected, when it happens, what you tried | |
| language | No | Optional language hint, e.g. python, javascript, elixir |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| fix | No | the concrete change to make | |
| model | No | ||
| usage | No | ||
| diagnosis | No | what is wrong, in plain language | |
| confidence | No | high, medium, or low | |
| fixed_code | No | corrected version of the snippet, when applicable | |
| root_cause | No |