How to Intermediate · 3 min read

How to use CrewAI with Ollama

Q: How to use CrewAI with Ollama

To use CrewAI with Ollama, you run Ollama models locally or on your server and interact with them via the ollama CLI or HTTP API. CrewAI can call these models by invoking the ollama command or HTTP endpoints, enabling seamless integration of Ollama's local AI models into CrewAI workflows.

Quick answer

To use CrewAI with Ollama, you run Ollama models locally or on your server and interact with them via the ollama CLI or HTTP API. CrewAI can call these models by invoking the ollama command or HTTP endpoints, enabling seamless integration of Ollama's local AI models into CrewAI workflows.

PREREQUISITES

Python 3.8+
Ollama installed and configured (https://ollama.com)
CrewAI account and API access
pip install requests

Setup Ollama and CrewAI

Install Ollama on your machine from https://ollama.com and ensure it is running. Install Python dependencies for making HTTP requests to Ollama's API or using its CLI. Also, have your CrewAI environment ready for integration.

bash

pip install requests

Step by step integration example

This example shows how to call an Ollama model from Python using the HTTP API and integrate the response into a CrewAI workflow.

python

import os
import requests

# Ollama local API endpoint
OLLAMA_API_URL = "http://localhost:11434"
MODEL_NAME = "llama2"

# Function to query Ollama model

def query_ollama(prompt: str) -> str:
    url = f"{OLLAMA_API_URL}/v1/chat/completions"
    headers = {"Content-Type": "application/json"}
    data = {
        "model": MODEL_NAME,
        "messages": [{"role": "user", "content": prompt}]
    }
    response = requests.post(url, json=data, headers=headers)
    response.raise_for_status()
    return response.json()["choices"][0]["message"]["content"]

# Example usage
if __name__ == "__main__":
    prompt = "Explain how CrewAI integrates with Ollama."
    answer = query_ollama(prompt)
    print("Ollama response:", answer)

    # Here you would send 'answer' to CrewAI API or use it in your CrewAI workflow

output

Ollama response: CrewAI can integrate with Ollama by invoking Ollama's local models via HTTP API or CLI, enabling seamless AI-powered workflows.

Common variations

Use the ollama CLI directly from Python with subprocess for local model calls.
Switch Ollama models by changing the MODEL_NAME variable.
Integrate asynchronously using httpx or aiohttp for non-blocking calls.

python

import subprocess

def query_ollama_cli(prompt: str) -> str:
    result = subprocess.run(
        ["ollama", "chat", "llama2", prompt],
        capture_output=True,
        text=True
    )
    return result.stdout.strip()

if __name__ == "__main__":
    print(query_ollama_cli("What is CrewAI?"))

output

CrewAI is a platform that enables AI workflow automation and integration with models like Ollama.

Troubleshooting

If you get connection errors, ensure Ollama daemon is running and listening on the correct port (default 11434).
Check your firewall or network settings if HTTP requests to Ollama fail.
Use ollama list CLI command to verify installed models.
For permission issues with CLI calls, run your Python script with appropriate user privileges.

✅

Key Takeaways

Use Ollama's local HTTP API or CLI to integrate with CrewAI workflows.
Switch Ollama models easily by changing the model parameter in API calls.
Ensure Ollama daemon is running to avoid connection errors.
Python's requests or subprocess modules enable flexible Ollama integration.
CrewAI can leverage Ollama models for powerful local AI processing.

Verified 2026-04 · llama2

Verify ↗