Troubleshooting

Something not working? Start here. Most issues fall into a handful of buckets, and almost all of them are fixed by checking three things first.

What to check first#

Before anything else, confirm these:

1. Base URL has /v1. The OpenAI-compatible base is https://api.fluxrouter.ai/v1, and the Anthropic-compatible base is https://api.fluxrouter.ai/anthropic. A missing /v1 or a bare /chat path is the single most common cause of connection failures. There is no bare /chat endpoint.
2. The key is your Flux key. It starts with sk-. Send it as Authorization: Bearer sk-... (the Anthropic path also accepts x-api-key: sk-...). A leftover OpenAI or Anthropic key will return 401.
3. The model is real. Use flux-auto (the default) or a real flux-* id. A model id your client rejects, or one Flux does not recognize, is the second most common failure. See Models.
4. Capture X-Flux-Request-Id. Every response includes it. When you ask for help, include that id so support can find your exact request.

Pages in this section#

• Error reference (401, 402, 429, 5xx): what each status code means and how to fix it.
• Truncated or empty output: responses that stop early or come back blank.
• Latency and timeouts: slow responses and client-side timeout errors.
• Find out which model answered and what it cost: read the X-Flux-* response headers.
• My tool won't connect: per-client gotchas for Codex, Cursor, Copilot, and others.

Still stuck?#

Grab the X-Flux-Request-Id from the failing response, note the exact base URL and model id you used, and contact support. Those three details resolve most tickets on the first reply.