@kilocode/cli runs the Kilo Code agent in your terminal. This page covers only the CLI; for the VS Code extension, see Kilo Code (VS Code).
The CLI speaks OpenAI-compatible APIs: point it at routing.run’s base URL and use route/… model IDs.
For the VS Code extension, see Kilo Code (VS Code).
rk_.
CLI
Install the CLI
Apply environment variables
Before running
kilo, use the CLI connection prompt below (or add the same exports to your shell profile).CLI connection prompt
The Kilo CLI expects OpenAI-compatible variables (OPENAI_API_KEY, OPENAI_BASE_URL). Paste this when onboarding yourself or a teammate. It includes the full integration flow.
Kilo Code CLI — connect to routing.run
Manual configuration
If you want to configure Kilo Code CLI manually instead of using the prompt above, export these exact values before you runkilo:
routing.run so switching providers is simpler.
Use a model id with the route/ prefix, for example route/glm-5.1.
Auth and flags: Kilo CLI docs. If the CLI reads its own config file, set the same base URL and key there.
Recommended models
Plan tier and access are authoritative on GEThttps://api.routing.run/v1/models with your rk_ key. The full list is on the models page; common picks:
| Model | Use case |
|---|---|
route/deepseek-v3.2 | General-purpose chat (matches most doc examples) |
route/glm-5.1 | Reasoning and coding |
route/minimax-m2.7 | Long-context sessions |
route/kimi-k2.5 | Agentic and tool-heavy flows |
route/qwen3.6-plus-preview | Qwen3.6-class preview |
route/qwen3.5-plus | Qwen3.5 flagship |