Quickstart

Use Ollama in desktop apps and coding agents, or build an application with an API key.

Get started

Download Ollama for macOS, Windows, or Linux. Open the app, or get started from your terminal:

ollama

Follow the setup prompts. Sign in to use cloud models, or choose a local model.

Connect a desktop app

On macOS, open Ollama and select Apps. Connect Claude or ChatGPT (Desktop). Follow the prompts to install or restart the app.

Choose your Ollama models in Settings → Apps.

• Claude Desktop — use Ollama models in Claude.
• ChatGPT Desktop — use Ollama models in Codex mode. Regular Chat and voice use your usual ChatGPT connection.

Use a coding agent

From your project directory, launch your agent:

ollama launch claude

Ollama offers to install Claude Code if needed. See Claude Code for details.

Install Codex CLI first, then launch it:

ollama launch codex

See Codex CLI for setup.

ollama launch opencode

Ollama offers to install OpenCode if needed. See OpenCode for details.

Choose a model when prompted, then try:

Explain how this repository is organized.

For more tools, see Integrations. To connect an agent directly with an API key, see Claude Code.

Build an application

Use cloud models with an API key, or run models locally without one. Cloud requests do not require an Ollama installation.

1. Create an API key

For local models, skip this step and select Local under Send a request.

For cloud models, sign in or create an account, then create an API key.

Set your key in the terminal:

export OLLAMA_API_KEY="your_api_key"

$env:OLLAMA_API_KEY = "your_api_key"

Keep your key on your application's server, outside browser code and source control.

2. Send a request

Make your first request to Ollama

These examples use the gemma4:31b cloud model.

curl https://ollama.com/api/chat \
  -H "Authorization: Bearer $OLLAMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma4:31b",
    "messages": [
      {
        "role": "user",
        "content": "Say hello in one sentence."
      }
    ],
    "stream": false
  }'

Read the answer from message.content. See Ollama's API and libraries.

curl https://ollama.com/v1/chat/completions \
  -H "Authorization: Bearer $OLLAMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma4:31b",
    "messages": [
      {
        "role": "user",
        "content": "Say hello in one sentence."
      }
    ]
  }'

Read the answer from choices[0].message.content. See OpenAI compatibility for client setup and supported features.

curl https://ollama.com/v1/responses \
  -H "Authorization: Bearer $OLLAMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma4:31b",
    "input": "Say hello in one sentence."
  }'

Read the output_text content blocks inside output. The OpenAI Python client exposes the text as response.output_text.

Responses requests are stateless: include conversation history in each request. previous_response_id and conversation aren't supported. See Responses compatibility.

curl https://ollama.com/v1/messages \
  -H "Authorization: Bearer $OLLAMA_API_KEY" \
  -H "Content-Type: application/json" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "gemma4:31b",
    "max_tokens": 1024,
    "messages": [
      {
        "role": "user",
        "content": "Say hello in one sentence."
      }
    ]
  }'

Read text blocks from content. Direct cloud requests use bearer authentication. See Anthropic compatibility for client setup and supported features.

OpenAI and Anthropic compatibility each cover a subset of the original API. See Cloud for models and usage limits.

Run Gemma 4 E2B on your computer. No API key required.

The model download is about 7.2 GB. We recommend 8 GB of available VRAM, or unified memory on a Mac. Larger context windows need more memory. With less VRAM, Ollama can use system RAM, but responses may be slower.

Download Ollama and open the app. On Linux, start the server with ollama serve if it is not already running.

Download the model:

ollama pull gemma4:e2b

Send a request to your local server:

curl http://localhost:11434/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma4:e2b",
    "messages": [
      {
        "role": "user",
        "content": "Say hello in one sentence."
      }
    ],
    "stream": false
  }'

curl http://localhost:11434/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma4:e2b",
    "messages": [
      {
        "role": "user",
        "content": "Say hello in one sentence."
      }
    ]
  }'

Read the answer from choices[0].message.content. See OpenAI compatibility for client setup and supported features.

curl http://localhost:11434/v1/responses \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma4:e2b",
    "input": "Say hello in one sentence."
  }'

Read the output_text content blocks inside output. The OpenAI Python client exposes the text as response.output_text.

Responses requests are stateless: include conversation history in each request. previous_response_id and conversation aren't supported. See Responses compatibility.

curl http://localhost:11434/v1/messages \
  -H "Content-Type: application/json" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "gemma4:e2b",
    "max_tokens": 1024,
    "messages": [
      {
        "role": "user",
        "content": "Say hello in one sentence."
      }
    ]
  }'

Read text blocks from content. See Anthropic compatibility for client setup and supported features.

Run a model locally

Download Ollama, then run:

ollama run gemma4:e2b

Ollama downloads the model and starts a chat on your computer. Type /bye to leave.

Next steps

Add tool calling, stream responses, or browse more models.