> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bondata.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Ollama

> Connect BonData to your self-hosted Ollama instance for local AI model inference

Ollama is an open-source tool for running large language models locally on your own hardware. With BonData, you can connect to a self-hosted Ollama instance to use local models in your Agent workflows - giving you full control over your data and infrastructure with no third-party API costs.

## Ollama Connection Setup

Follow these steps to connect BonData to your Ollama server.

### Step 1: Ensure Ollama is Accessible

Your Ollama server must be reachable from the internet. By default, Ollama runs on port `11434`.

1. Verify your server is running and accessible at its public URL
2. Test the connection by visiting `https://your-server.com:11434/api/tags` - you should see a JSON response listing available models

### Step 2: Enter Connection Details

Enter your Ollama server URL in BonData (e.g., `https://ollama.your-server.com:11434`).

<Warning>
  Your Ollama server must have a public URL reachable from the internet. Local-only instances (e.g., `localhost`) will not work.
</Warning>

<Note>
  No API key is required - Ollama uses the server URL for authentication.
</Note>

## Required Fields

| Field    | Description                                                                     |
| -------- | ------------------------------------------------------------------------------- |
| Host URL | Public URL of your Ollama server (e.g., `https://ollama.your-server.com:11434`) |