Connect Warp Terminal’s Agent to Ollama API

Warp’s intelligent terminal now lets you bring external AI providers into its built-in agent. One of the most useful options is Ollama—whether you’re running models locally or using Ollama’s cloud endpoints—so you can keep coding with private, low-cost, or free models without leaving the terminal.

Official documentation can be sparse on the exact base URL, which often leads to connection errors. The reliable method is to use Ollama’s OpenAI-compatible endpoint (the one that ends with /v1). In Warp’s agent settings, add a custom endpoint, paste the correct base URL, enter your Ollama API key, give the endpoint a clear name, and select the model you want (for example a recent MiniMax or other available model).

Once configured, switch into agent mode and ask a simple verification question such as “What model are you?” and “What are your coding capabilities?” A successful reply confirms the connection is live. You can then use the full agent workflow powered by your chosen Ollama model.

This setup gives you the polished Warp experience—rich input, agent features, and terminal integration—while routing requests through Ollama instead of the default cloud providers. It’s ideal for privacy, cost control, or experimenting with the latest open models.