How to Configure a Custom API Host in Chatbox

Chatbox is a cross-platform AI client that can connect to different model services. If you use your own API key, the setup usually comes down to API Host, API Key, and model name.

Check your Chatbox version first

Field names can vary by version. You may see:

• API Host
• API URL
• Base URL
• OpenAI API Host
• Custom Provider

Look at the example next to the field. If it looks like https://api.openai.com/v1, use a URL with /v1. If it looks like https://api.openai.com, use the root URL.

Recommended setup

If your Chatbox version asks for an OpenAI-compatible endpoint, start with:

API Host / Base URL: https://api.cheng-zi-ai.com/v1
API Key: sk-your-api-key
Model: gpt-5.3-codex

If you see a 404 or route-related error, try removing /v1:

API Host / Base URL: https://api.cheng-zi-ai.com

Setup steps

1. Open Chatbox settings.
2. Find model service, AI provider, or API settings.
3. Choose OpenAI or a custom OpenAI-compatible service.
4. Paste your API Key.
5. Enter the API Host or Base URL.
6. Enter or select a model name.
7. Save and send a short test prompt.

Test prompt

Reply with exactly: OK

If it returns OK, switch to the model you actually want to use. Avoid debugging with expensive models first.

Common issues

401

Check whether the API Key was copied completely, whether it has extra spaces, and whether it was pasted under the right provider.

404

The API Host path probably does not match what the client expects. Test whether /v1 should be included.

Model not found

Copy the Model ID from the console instead of typing it manually. Some clients separate display names from actual model IDs.

When to test another client

If your Chatbox version cannot reliably save a custom API Host, test the same key in Cline, Cherry Studio, or NextChat. If it works elsewhere, the key itself is fine.