Ollama OpenAI Compatible API

Ollama ships an OpenAI compatible API under http://localhost:11434/v1. Many tools and libraries were written for the OpenAI Chat Completions format; by changing only the base URL you can point them at local models.

Chat Completions with curl

curl http://localhost:11434/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama3.2",
    "messages": [{ "role": "user", "content": "Say hello in Spanish." }]
  }'

{
  "id": "chatcmpl-123",
  "object": "chat.completion",
  "model": "llama3.2",
  "choices": [{
    "index": 0,
    "message": { "role": "assistant", "content": "¡Hola!" },
    "finish_reason": "stop"
  }],
  "usage": { "prompt_tokens": 30, "completion_tokens": 4, "total_tokens": 34 }
}
(Sample response.)

Using the OpenAI Python SDK

An API key is required by the SDK but ignored by Ollama, so any string works.

from openai import OpenAI

client = OpenAI(base_url="http://localhost:11434/v1", api_key="ollama")

resp = client.chat.completions.create(
    model="llama3.2",
    messages=[{"role": "user", "content": "Give me one tip for clean code."}],
)
print(resp.choices[0].message.content)

Use clear, descriptive names for variables and functions so the code explains itself.
(Sample response.)

Using the OpenAI JavaScript SDK

import OpenAI from 'openai';

const client = new OpenAI({ baseURL: 'http://localhost:11434/v1', apiKey: 'ollama' });

const resp = await client.chat.completions.create({
  model: 'llama3.2',
  messages: [{ role: 'user', content: 'What is 12 * 8?' }],
});
console.log(resp.choices[0].message.content);

12 * 8 = 96.
(Sample response.)

Supported Endpoints

At the time of writing (2026) Ollama supports the most common OpenAI-style endpoints, including:

/v1/chat/completions - chat, streaming, tools, JSON mode, images
/v1/completions - plain text completion
/v1/embeddings - vector embeddings
/v1/models - list local models

Compatibility is good but not perfect: provider-specific features may be missing. When you need everything Ollama offers (keep_alive, raw options) use the native /api endpoints instead.

Frequently asked questions

Do I need an API key for Ollama?

No. The local server has no authentication. OpenAI SDKs require a key value, so pass any placeholder like "ollama".

Can I switch between OpenAI and Ollama easily?

Yes. If your code uses the Chat Completions format, change base_url and model name, ideally through environment variables.

Does the Ollama OpenAI endpoint support streaming?

Yes, set stream: true exactly as you would with the OpenAI API.