Ollama · REST API
Use the OpenAI Python client
Point the OpenAI client at your local server. The API key is required but ignored.
from openai import OpenAI
client = OpenAI(base_url="http://localhost:11434/v1", api_key="ollama")
reply = client.chat.completions.create(model="llama3.2", messages=[{"role": "user", "content": "Hello"}])