Troubleshooting
All errors are explicit — we never silently fall back or swap your model. Match the error code below to the fix.
model_not_available (403)
The model isn't in your group. Ask an operator to add it to your group, or use a model that GET /v1/models
returns for your key.
model_not_found (404)
The model doesn't support the protocol you used (e.g. you called an Anthropic-only model via
/v1/chat/completions). Use the protocol the model supports — check Models & pricing.
insufficient_balance (402)
Your balance reached zero. Top up to resume; the cabinet and history stay available.
too_many_keys (400)
You hit the maximum number of active API keys for your group. Disable an old key (a disabled key frees a slot) or ask for a higher limit.
protocol_not_allowed / model_not_allowed (403)
The key itself restricts protocols or models. Create/use a key whose allowlist includes what you need.
endpoint_not_configured (502)
The model has no active endpoint for that protocol right now. Try the other protocol or contact support.
401 Unauthorized
Missing or wrong key. Send Authorization: Bearer YOUR_API_KEY; the raw key is shown only once at creation —
create a new one if you lost it.
Wrong base URL / path
- OpenAI clients: base URL
https://ai-api.cloud/v1, path/chat/completions. - Anthropic clients: base URL
https://ai-api.cloud, path/v1/messages(+anthropic-versionheader).