Connection Troubleshooting

Work through these checks in order. Change one setting at a time.

1. Confirm the API key

  • Remove leading or trailing spaces.
  • Confirm the key is enabled, unexpired, and has remaining quota.
  • Confirm the process can read the configured environment variable.
  • Do not paste the full key into logs or support messages.

2. Confirm the protocol and URL

ProtocolBase URLOperation added by client
Chat Completionshttps://getrouter.ai/v1/chat/completions
Responseshttps://getrouter.ai/v1/responses
Claude Messageshttps://getrouter.ai/v1/messages
Geminihttps://getrouter.ai/v1beta/models/...

A 404 often means the client added /v1 or an operation path twice.

3. Confirm the model

  • Copy the exact model ID from Models.
  • Confirm the key is allowed to use that model.
  • Confirm the model supports the protocol and feature being requested.
  • Enter the model manually if the tool cannot load a model list.

4. Interpret common statuses

StatusMeaning to check first
400Unsupported field, invalid request, or feature mismatch
401Missing or invalid API key
403Key, model, group, or policy restriction
404Wrong URL/path or unknown model
429Balance, quota, concurrency, or rate limit
5xxTemporary upstream or gateway failure

5. Test the minimum request

Disable tools, files, images, long context, JSON mode, and custom reasoning settings. Send one short text prompt. Add advanced features back one at a time.

Ask for help safely

Provide the tool/version, protocol, model ID, redacted Base URL configuration, HTTP status, error text, Request ID, and time with timezone. Never provide the complete API key.