Unlock Powerful AI in Your Terminal: Connect OpenRouter to Warp CLI via BYOK/Custom Endpoint

Warp Terminal has made a game-changing update: full support for Bring Your Own Key (BYOK) and custom OpenAI-compatible inference endpoints — now available even on the free plan. This means you can supercharge Warp’s AI agents with hundreds of models from OpenRouter, including strong coding, reasoning, and general-purpose LLMs, without relying solely on Warp’s built-in options.

OpenRouter acts as a unified gateway to top models (Claude, DeepSeek, Kimi, Gemini, Llama, Qwen, and many more) at competitive prices, with plenty of free options to test. By connecting it to Warp, you get seamless AI assistance directly in your terminal for coding, debugging, scripting, command generation, and agentic workflows.

Step-by-Step Setup (Super Simple)

1. Open Warp Terminal and go to Settings → Agents → Custom Inference.
2. Click Add Custom Endpoint.
3. Set the name (e.g., “OpenRouter”).
4. Base URL: https://openrouter.ai/api/v1
5. Paste your OpenRouter API key.
6. Add a model ID (e.g., minimax/minimax-m2-5 or any preferred coding model).
7. Save and start a new session.

Once connected, press the shortcut (usually Ctrl + I) to invoke AI, switch models easily, and enjoy full access to OpenRouter’s vast model library inside Warp.

This setup is perfect for developers who want flexibility, cost control, and access to the latest open-source and proprietary models without switching tools. Whether you’re on free or paid, it dramatically expands what Warp’s AI can do.