# Use Parallel with Ollama

Give local Ollama models access to the live web by exposing Parallel as a callable tool.

Add Parallel to Ollama[Add Parallel to Ollama]Get an API key[Get an API key]

## What you can do

- Combine local models with current web search and research.
- Let an Ollama model invoke Parallel only when a prompt needs live information.

## How it works

Expose Parallel as a callable tool through Ollama's tool-calling interface: the local model emits a tool call when a prompt needs live information, your code executes the Parallel request, and the sourced result goes back to the model for its final response.

Local inference stays local; only the web research round-trips to Parallel.

## Set up in three steps

1. Define the Parallel operation as a tool available to the Ollama model.
2. Map the tool arguments to a Parallel request and configure credentials.
3. Execute requested calls and return the sourced result to the model for its final response.

Add Parallel to Ollama[Add Parallel to Ollama]