add-ollama-provider

Route a NanoClaw agent group to a local Ollama model by applying per-agent group environment variable overrides and setting the Claude model configuration. Ollama natively supports the Anthropic API, so no provider code changes are needed.

Yes, Ollama speaks the Anthropic API natively, allowing you to run agents locally. You can redirect API calls to a local Ollama instance by configuring environment variables and the model setting without altering provider code.

Before switching to an Ollama local model, you must install Ollama, run the selected model, and configure per-agent group environment overrides and host blocking. You also need to set the specific model setting for the agent group.

Use a local Ollama model to reduce API costs and enable experimentation with open-weight models. It supports running agents locally across development and testing scenarios without incurring external API expenses.

No, redirecting to Ollama does not require provider code changes because Ollama natively supports the Anthropic API. You only need to apply environment variable overrides and update the model setting for the group.

Using local Ollama models requires host blocking configuration and per-agent group environment overrides, which limits changes to specific groups. You must also ensure Ollama is installed and running before executing agents.