Skip to main content
OpenClaw is a personal AI assistant that runs on your own devices. It bridges messaging services like WhatsApp, Telegram, Slack, Discord, and iMessage to AI models through a centralized gateway.
Previously known as Clawdbot. The alias ollama launch clawdbot still works.

​
Quick Start

ollama launch openclaw
Ollama handles everything automatically:
1

Installation

If OpenClaw isn’t installed, Ollama prompts to install it via npm
2

Security Notice

On first launch, a security notice explains the risks of tool access
3

Model Selection

Pick a model from the interactive selector (local or cloud)
4

Onboarding

Ollama configures the provider, installs the gateway daemon, and sets your model as primary
5

Launch

The gateway starts in the background and the OpenClaw TUI opens

​
Features

100+ Skills

Built-in skills for file management, web search, code execution, and more

Multi-Platform

Connect WhatsApp, Telegram, Slack, Discord, iMessage, and other messaging apps

Local Gateway

Runs a secure gateway on your machine for private AI conversations

Plugin System

Extensible with custom plugins and skills

​
Configuration

​
Change Model Without Launching

ollama launch openclaw --config

​
Use a Specific Model

ollama launch openclaw --model kimi-k2.5:cloud
If the gateway is already running, it restarts automatically to pick up the new model.

​
Pass Custom Arguments

ollama launch openclaw -- gateway run --port 8080
Arguments after -- are passed directly to the openclaw binary.
OpenClaw requires a large context window (at least 64k tokens) to handle complex tool calling and multi-turn conversations.

​
Cloud Models

kimi-k2.5:cloud

Multimodal reasoning with subagents (260k context)

minimax-m2.5:cloud

Fast, efficient coding and productivity (200k context)

glm-5:cloud

Reasoning and code generation (200k context)
More cloud models at ollama.com/search?c=cloud.

​
Local Models

glm-4.7-flash

Reasoning and code generation locally (~25 GB VRAM)
See Context Length for configuration.

​
Connect Messaging Apps

After launching OpenClaw, connect your messaging platforms: 
openclaw configure --section channels
This opens an interactive wizard to link:
  • WhatsApp
  • Telegram
  • Slack
  • Discord
  • iMessage (macOS only)
  • SMS (with supported providers)

​
Security Considerations

OpenClaw can read files and execute commands when tools are enabled. A malicious prompt could potentially cause it to perform unsafe actions.
On first launch, Ollama displays a security notice. OpenClaw:
  • Has access to files in your home directory
  • Can execute shell commands when tools are enabled
  • Runs a local gateway that accepts connections from configured messaging apps
  • Stores configuration and credentials locally
Learn more: OpenClaw Security Documentation

​
Gateway Management

The OpenClaw gateway runs in the background to handle requests from messaging apps.

​
Check Status

openclaw gateway status

​
Stop the Gateway

openclaw gateway stop

​
Restart the Gateway

openclaw gateway restart

​
View Logs

openclaw gateway logs

​
Skills and Plugins

​
Browse Available Skills

openclaw skills

​
Install a Skill

openclaw skills install <skill-name>

​
Web Search Plugin

When using cloud models, Ollama automatically installs the web search plugin for enhanced capabilities:
  • Real-time web searches
  • News and information retrieval
  • Fact-checking and verification

​
Configuration Files

OpenClaw stores its configuration in ~/.openclaw/openclaw.json. 
{
  "models": {
    "providers": {
      "ollama": {
        "baseUrl": "http://localhost:11434/v1",
        "apiKey": "ollama-local",
        "api": "ollama",
        "models": [
          {
            "id": "kimi-k2.5:cloud",
            "name": "kimi-k2.5:cloud",
            "input": ["text", "image"],
            "reasoning": true,
            "contextWindow": 262144,
            "maxTokens": 262144,
            "cost": {
              "input": 0,
              "output": 0,
              "cacheRead": 0,
              "cacheWrite": 0
            }
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "ollama/kimi-k2.5:cloud"
      }
    }
  },
  "gateway": {
    "port": 18789,
    "auth": {
      "token": "your-token-here"
    }
  }
}
Ollama automatically manages this file when you use ollama launch openclaw.

​
Web UI

OpenClaw includes a web interface for managing conversations and settings: 
http://localhost:18789
The URL includes your auth token automatically when launched via ollama launch.

​
Troubleshooting

​
OpenClaw Won’t Install

OpenClaw requires Node.js and npm:
  1. Install Node.js from nodejs.org
  2. Verify installation: npm --version
  3. Rerun: ollama launch openclaw

​
Gateway Won’t Start

Check if the port is already in use: 
lsof -i :18789  # macOS/Linux
netstat -ano | findstr :18789  # Windows

​
Model Not Found

Ensure the model is pulled: 
ollama pull kimi-k2.5:cloud
ollama list

​
Windows Support

OpenClaw works best on WSL2 (Windows Subsystem for Linux): 
wsl --install
See: OpenClaw Windows Guide

​
Advanced Usage

​
Custom Gateway Port

Edit ~/.openclaw/openclaw.json to change the gateway port: 
{
  "gateway": {
    "port": 8080
  }
}

​
Multiple Models

Configure different models for different use cases: 
ollama launch openclaw --model qwen3:8b
The most recently selected model becomes the primary.

​
Direct CLI Usage

Bypass the launcher for advanced control: 
export OLLAMA_HOST=http://localhost:11434
openclaw onboard --non-interactive
openclaw gateway run
openclaw tui

​
Learn More

Official Docs

Complete OpenClaw documentation

npm Package

OpenClaw on npm

Skills

Browse available skills

Security

Security best practices