Skip to main content

​
Start with the request path

First check whether the client can reach AIOHub, and whether the API key, group, and protocol match the model you are calling.
  • Connection fails before AIOHub responds: check the client Base URL, local proxy, firewall, and API key environment variables.
  • AIOHub returns an error: use the status code, model name, group, request ID, and error message to continue.
  • The client shows a different error: check whether the client failed to parse the response or retried the operation and surfaced the later error.

​
Common issues

​
Base URL includes the wrong path

Do not put a full endpoint in a client Base URL field. If an error path contains /v1/v1, /chat/completions/chat/completions, or a repeated /v1beta, the client configuration probably includes an extra path segment. Check the correct protocol-specific Base URL in Choose a client.

​
400 errors

These usually mean the request body, parameters, or protocol do not match. Confirm the endpoint and request body belong to the same API format, such as OpenAI Chat Completions, OpenAI Responses, Claude Messages, or Gemini API. For client-session-related 400 errors, check the matching client page.

​
401 or 403 errors

401 usually means the API key is invalid, expired, disabled, or sent in the wrong header. 403 usually means the current group cannot access the requested model or endpoint. Check that the API key is enabled, the account has balance, the group contains the target model, and the client page uses the right protocol and Base URL.

​
API Connect Error

  • Check the local network connection.
  • Check proxy settings.
  • Confirm the client Base URL does not contain a repeated path.
  • Request the model list to validate API key and network access:
curl https://api.aiohub.org/v1/models \
  -H "Authorization: Bearer sk-your-api-key"

​
Request timeout

  • Check network latency.
  • Some models respond more slowly under load.
  • Reduce input size or switch models, then retry.

​
Model not available

Model availability depends on the current service state, the groups assigned to the API key, and the endpoint used by the client. Check the console pricing page, or request /v1/models or /v1beta/models to confirm the models visible to the current API key.

​
Get help

If these checks do not resolve the issue, collect the request time, request ID, model name, API key name, client name, Base URL, and error message. Then check FAQ and console announcements.