Chatbox
Chatbox supports custom model providers. Use OpenAI-compatible mode; you do not need a built-in gateway brand entry.
1. Add a provider
Open Settings → Model Provider → Add. Name it Fenno and choose OpenAI API Compatible.
2. Enter the address and key
Some versions expose separate Host and Path fields; others use a single Base URL. Follow the fields shown in your client:
| Field | Value |
|---|---|
| API Host, when separate from Path | https://api.fenno.ai |
| API Path, when configured separately | /v1/chat/completions |
| Base URL, when no separate Path is provided | https://api.fenno.ai/v1 |
| API key | Your Fenno API key |
The final chat request must target https://api.fenno.ai/v1/chat/completions. For a 404, check the assembled URL instead of adding /v1 to every field.
3. Add a model
Fetch the model list and add the models you need, or enter an exact model ID from the console manually. Display names are customizable; the ID sent to the API must match.
Save, then choose a model under Fenno in the conversation’s model picker. Do not leave the previous default provider selected.
4. Test the connection
Run Check, then start a new conversation with “Reply with OK only”. Match the request time and model in the Fenno console to verify the provider actually used.
Working text chat does not guarantee image generation, web search or tool calling. Those features also depend on the Chatbox version, protocol and model. Disable extras while diagnosing basic connectivity.
Troubleshooting
| Symptom | Action |
|---|---|
| 404 | Check for duplicated /v1 or /chat/completions in Host and Path |
| 401 | Check the key source and full value, then save the provider again |
| Model missing from the picker | Add and enable it, using an exact ID if discovery fails |
| Check succeeds but chat fails | Check the conversation’s provider, model and extra feature settings |
| Only older conversations work | Test a new conversation without inherited model settings |
See the official Chatbox custom provider guide .