Tapix AI

Troubleshooting

First run the "One-click Connectivity Test" below; most issues can be located on your own. If it still fails after testing, just paste the output to us.

One-click connectivity test

Paste your API Key (sk-fg-…), we test the gateway once with a direct connection from your browser.

What to do about common status codes

401
Key is invalid or has been disabled.
Go to Console → API keys Copy the latest one; note that the "prefix" is not the full key (we only show the full value once).
402
Insufficient balance or quota exhausted.
Go to Add credit; after topping up, the balance is credited immediately, no restart needed.
403
This key is bound to other models, or the model is not available for your group.
Check the models available on your key, or create an all-models key.
404
Model name does not exist.
See Model name quick reference copy the exact name, or call GET /v1/models for the list.
429
Rate limit triggered (default: 60 requests/minute, 200,000 tokens/minute).
By response header retry-after seconds of backoff before retrying; if it's not enough long-term, contact us to raise the limit.
5xx
Upstream error on our side.
The gateway will automatically switch to a backup channel; if it keeps happening, send us the timestamps (accurate to the minute) and we'll check the logs.

Three self-check actions

  1. Confirm base_url is https://YOUR_HOST/v1 (do not add a second /v1).
  2. Confirm the request headers are Authorization: Bearer sk-fg-…, with a space in between.
  3. Confirm model The field is the model name, not the display name (case-insensitive, but don't write it with spaces like "Claude Opus 5.5").