Debugging requests - Nito

Two things diagnose a failure: X-Request-ID on the response and code in the body. Log both on every call. When you contact support, include X-Request-ID and, when a request failed after dispatch, nito_operation_id from the body.

curl -i "https://app.trynito.ai/v1/chat/completions" \
  -H "Authorization: Bearer z_sk_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{ "model": "z-ai/glm-5.2", "messages": [{ "role": "user", "content": "Hi" }] }'
# look for: X-Request-ID: <value>

Branch on code, then read message and param for detail.

By Status

Status	Likely cause	What to check
`400`	Invalid request, unsupported feature/attachment, or unresolved model (`param: model`).	Validate the JSON body, the `model`, and any attachments. Do not retry unchanged.
`401`	Missing, malformed, revoked, or expired key.	The key starts with `z_sk_` and is sent as `Authorization: Bearer`.
`404`	Wrong path (`unknown_route`).	The base URL ends in `/v1` for OpenAI-style clients.
`429`	Rate limit or quota.	The `X-RateLimit-*` headers, `Retry-After`, and whether `code` is `rate_limited` or `insufficient_quota`.

For 400/401/403, fix the request or credential. For 429/502/503, retry with backoff. If it persists, send support@trynito.ai the X-Request-ID, the code, the nito_operation_id if present, and the timestamp.