Tools and compatibility
Connecting a tool to LYRouter comes down to the protocol, the base URL, the gateway API key and the model identifier. Each tool page starts with installation prerequisites and installation steps; if you have already installed the tool, skip straight to the configuration section.
Choose a guide
| Tool | Main way to connect |
|---|---|
| Claude Code | Anthropic Messages; the Base URL is the gateway root address |
| Codex CLI | Responses; the Base URL is the gateway address plus /v1 |
| ChatGPT | The desktop app works locally; configure a Responses model service and API key |
| OpenCode | Custom provider; the protocol follows the adapter package you choose |
| TRAE | Custom OpenAI-compatible model |
| WorkBuddy | Custom model; tell the Base URL apart from the full endpoint |
| CC Switch | Manages tool configurations; the address depends on the target tool's protocol |
| Other OpenAI-compatible clients | Cherry Studio, Chatbox, Cline and others; add an OpenAI-type provider, and the Base URL is the gateway address plus /v1 |
Prerequisites for installing and calling
Installing a tool may require a particular operating system, runtime environment, permissions and a tool account; calling LYRouter requires an enabled gateway API key, an available model and a valid service source. Confirm the two sets of conditions separately.
How to verify
Run a minimal text request first, then try tool calls. Generation requests sent by the client should appear in the LYRouter request logs; fetching the model list or only testing the connection does not show that generation works.