Other Clients & Protocols
Chatbox
Configure the existing OpenAI or Claude provider in Chatbox and troubleshoot connection issues.
Configure the existing OpenAI or Claude provider in Chatbox with models for your current key.
1Update the OpenAI or Claude API Host
Open Settings → Model Provider, select the existing OpenAI or Claude provider, update its API Host and enter your key. No custom provider is needed.
OpenAI API
YOUR_API_KEY in the examples is a placeholder and will not work as copied. Sign in to pick one of your keys and fill it into every example on this page.
Sign in to fill in keyhttps://api.keyvia.cc/v1YOUR_API_KEYReplace this with an API key from the console, without a Bearer prefix, spaces or line breaks.
Set API Host to the address above, including /v1. Do not append /chat/completions or use the website URL or the /api/v1 management API.
Anthropic API
https://api.keyvia.cc/v1YOUR_API_KEYReplace this with an API key from the console, without a Bearer prefix, spaces or line breaks.
In Claude mode, set API Host to https://api.keyvia.cc/v1. The /v1 suffix is required; the connection will not work without it.
2Add a model and select the conversation provider
Copy the full model ID for your key. Add or select it under the chosen OpenAI or Claude provider, save, then select the same provider and model in your conversation. API Keys · Models
A failed model-list fetch does not necessarily mean chat requests will fail. Try adding an available model manually. Existing conversations may still use the previous provider. Start with plain text and leave images, web search and other tools disabled for the first check.
3Verify and troubleshoot
Save, run the connection check, and send a short message. Check the request in your console usage records. Real chat requests incur normal usage charges.
- 401: check that the key is valid and complete. 403: inspect the error body, group permissions and origin restrictions rather than assuming the key is wrong.
- 404: the final path should be /v1/chat/completions for OpenAI API or /v1/messages for Anthropic API. For model-not-found errors, verify the complete model ID against the models available to your key.
- For Failed to fetch or network errors, first confirm that API Host includes /v1, then check connectivity and CORS preflight. Share the app version, error and request time with support, never your key.