Skip to main content
Ollama provides a powerful launcher system that lets you quickly connect local and cloud models to coding agents, IDE extensions, and productivity tools.

​
The Launch System

The ollama launch command provides an interactive way to configure and run integrations: 
ollama launch
This opens an interactive menu where you can:
1

Select an integration

Choose from coding agents, IDEs, or assistants
2

Pick a model

Select from your local models or cloud models with large context windows
3

Launch instantly

Ollama handles configuration and starts the integration

​
Direct Launch

You can also launch specific integrations directly: 
ollama launch claude
ollama launch opencode
ollama launch openclaw

​
Configuration Options

# Interactive setup and launch
ollama launch claude

​
Coding Agents

Autonomous coding assistants that can read, modify, and execute code:

Claude Code

Anthropic’s coding tool with subagents

Codex

OpenAI’s open-source coding agent

OpenCode

Anomaly’s open-source terminal coding agent

Droid

Factory’s coding agent for terminal and IDEs

​
Key Features

  • Multi-file editing — Make changes across your entire codebase
  • Tool calling — Execute shell commands and run tests
  • Context-aware — Understands project structure and dependencies
  • Subagents — Some tools support delegating tasks to specialized agents
Coding agents work best with models that have at least 64k token context windows. See Context Length for configuration.

​
Personal AI Assistant

OpenClaw

Personal AI with 100+ skills across WhatsApp, Telegram, Slack, Discord, and more
OpenClaw bridges messaging platforms to AI models:
  • Connect your favorite messaging apps
  • Access 100+ built-in skills and plugins
  • Runs a local gateway for secure, private AI
  • Supports both local and cloud models

​
IDE Extensions

Native integrations for popular development environments:

VS Code

Microsoft’s popular code editor

JetBrains

IntelliJ, PyCharm, and other JetBrains IDEs

Zed

High-performance collaborative editor

Cline

VS Code extension for autonomous coding
See the IDE Extensions page for setup instructions.

​
Official Libraries

Official SDKs for Python and JavaScript make it easy to integrate Ollama into your applications:
  • ollama-python — Python library with async support
  • ollama-js — JavaScript/TypeScript library for Node and browsers
See the Libraries page for usage examples and community SDKs.

​
How Launch Works

The ollama launch system automatically:
  1. Detects installation — Checks if the integration binary is available
  2. Offers installation — For tools like OpenClaw, offers to install via npm
  3. Manages configuration — Updates config files with your selected models
  4. Handles authentication — Signs you in for cloud models if needed
  5. Syncs settings — Keeps your preferences across launches

​
Saved Configuration

Ollama saves your model preferences per integration: 
# First time
ollama launch claude
# → Interactive model selection

# Next time
ollama launch claude
# → Uses your saved model

​
Model Requirements

Most coding integrations need large context windows (64k-128k tokens) to understand entire codebases.
Recommended cloud models:
  • minimax-m2.5:cloud — Fast, efficient coding (200k context)
  • glm-5:cloud — Reasoning and code generation (200k context)
  • kimi-k2.5:cloud — Multimodal reasoning with subagents (260k context)
  • qwen3-coder:480b-cloud — Advanced code generation (260k context)
Recommended local models:
  • glm-4.7-flash — Local reasoning model (~25GB VRAM)
  • qwen3:8b — Efficient all-purpose assistant (~11GB VRAM)

​
Aliases and Model Routing

Some integrations like Claude Code and Codex support model aliases for routing requests:
  • primary — Default model for all requests
  • fast — Lightweight model for quick tasks
  • Custom aliases — Route specific agent types to different models
Ollama configures these automatically during launch.

​
Environment Variables

Integrations respect these environment variables:
  • OLLAMA_HOST — Ollama server URL (default: http://localhost:11434)
  • OLLAMA_API_KEY — API key when connecting to ollama.com

​
Community Integrations

Ollama works with hundreds of community tools. See the main README for:
  • Chat interfaces and web UIs
  • Frameworks and agent platforms
  • RAG and knowledge bases
  • Productivity apps
  • Development tools

​
Next Steps

Try a Coding Agent

Get started with OpenCode in your terminal

Setup IDE Extension

Configure VS Code or JetBrains

Install OpenClaw

Connect messaging apps to AI

Use Libraries

Build with Python or JavaScript SDKs