Ollama

Overview

The Ollama integration lets your agents use local LLMs running on your own hardware. This is ideal for privacy-sensitive tasks, reducing API costs, or using specialized open-source models.

Setup

1. Install Ollama on your server from ollama.ai
2. Pull the models you want to use: ollama pull llama3
3. Enter your Ollama endpoint URL in the Integrations page
4. Assign Ollama models to specific agents based on their needs

Best Practices

Assign small, fast local models for routine tasks (email classification, simple analysis) and reserve cloud models (Claude, GPT-4) for complex reasoning. This optimizes both cost and performance.