Chat clients
Desktop and web chat clients support custom OpenAI-compatible endpoints. Fill in the endpoint, key and model name and you are ready.
One-click import
In Console → Tokens, click the “Chat” button to the right of a token to import the configuration directly into installed clients such as Cherry Studio, CC Switch, DeepChat and Lobe Chat, with no manual entry.
General settings
| Field | Value |
|---|---|
| Provider type | OpenAI or OpenAI Compatible (for Claude models you can also choose Anthropic) |
| API address / Base URL | https://api.tokilabs.xyz/v1Some clients add /v1 automatically; in that case enter https://api.tokilabs.xyz |
| API Key | sk-your-key |
| Model | Add manually, e.g. claude-sonnet-5-5, gpt-6.1-sol |
Cherry Studio
- Settings → Model Provider → click “Add” at the bottom, and choose type
OpenAI(GPT) orAnthropic(Claude). - Enter
https://api.tokilabs.xyzas the API address; Cherry Studio completes the path automatically. - Enter your API key, click “Manage” to add models from the list, then click “Check” to test the connection.
Chatbox
- Settings → Model Provider → add a custom provider, and set API mode to
OpenAI API Compatible. - Enter
https://api.tokilabs.xyzas the API host and keep the API path as/v1/chat/completions. - Enter your key and add model names.
LobeChat / NextChat
In the OpenAI provider settings, enter the API proxy address https://api.tokilabs.xyz/v1 (for NextChat, https://api.tokilabs.xyz) and the API Key, then add the model names you need under custom models.
CC Switch
CC Switch lets you switch between multiple Claude Code / Codex providers. When adding a provider, use https://api.tokilabs.xyz for Claude Code and https://api.tokilabs.xyz/v1 for Codex, with the same key for both.