Using Custom LLM Providers with the Agent

While our agent comes pre-configured to work with OpenAI and Anthropic, you can integrate custom LLM providers in two ways:

Method 1: Using OpenAI-Compatible Endpoints

If your custom LLM framework can produce OpenAI-compatible responses (following the `/chat/completions` format), you can use it by:

1. Setting up your framework to expose an HTTP endpoint that matches OpenAI's format

2. Configuring the agent to use your custom endpoint by changing the base_url in the OpenAI LLM settings

3. Optionally adding any required custom headers or parameters

Method 2: Customizing the LLM Node

For more advanced integrations, you can customize the `llm_node` directly to call your LLM provider. This gives you complete control over how the agent interacts with your LLM.

You can find an example of customizing the LLM node in our GitHub repository.