🦞 OpenClaw + 🦙 Ollama Installation Guide

中文版 | English | 網頁版 | Web-site Version

📰 Latest OpenClaw Battle Status Daily Updates: Highly recommended to read the OpenClaw Universe Battlefield Observation Log 🦞 before installation.

⚠️ Important Note: Native Windows vs WSL2

You are currently viewing the Native Windows Installation Guide. This method allows you to quickly experience OpenClaw + Ollama, but has the following limitations:

Feature	Native Windows	WSL2 Version
Basic Chat	✅ Fully Supported	✅ Fully Supported
Memory Feature	⚠️ May be unstable	✅ Fully Supported
Skills Extension	⚠️ Only some Windows-compatible skills work	✅ Most skills supported
Homebrew Dependencies	❌ Not Supported	✅ Optionally Supported

Recommendations: - If you just want a quick experience of OpenClaw + Ollama → Continue reading the Native Windows Guide below. - If you need full features (memory, skills) → Please use the WSL2 Installation Guide. - Installed the Windows version but encountered issues → See Migrating to WSL2.

📚 More info: Why do we need WSL2?

A complete step-by-step guide for quickly installing OpenClaw and a local LLM (Ollama) natively on Windows.

⚠️ Version Requirements: Ollama v0.15.4+ and OpenClaw 2026.2.5+

📋 Table of Contents

Installation Steps

Environment Preparation
Install Ollama (v0.15.4+)
Install Python
Ollama Model Configuration
Recommended Models
Pull Local Models
Configure Cloud Models (Optional)
OpenClaw Installation
Install OpenClaw
Initial Configuration
9. Start Gateway Service
10. Configure OpenClaw to use Ollama
11. Test Gateway
Advanced Configuration
Telegram Bot Setup
Pair Telegram Channel
Other Advanced Settings (Optional)

1️⃣ Environment Preparation

Install Ollama (v0.15.4+)

⚠️ Important: Please make sure to install v0.15.4 or above, this version supports OpenClaw native integration.

Method 1: Using winget (Recommended)

winget install ollama

Method 2: Manual Download

Go to https://ollama.com/ to download the latest Windows executable and install it.

Verify Installation

ollama -v

Install Python

OpenClaw's Windows version does not automatically install Python, but many tasks require it:

winget install python

2️⃣ Ollama Model Configuration

Recommended Models

Although OpenClaw theoretically supports any OpenAI-compatible model, community and official tests show the following models perform better:

Model Series	Model Name	VRAM Req	Size	Best For
GLM	`glm-4.7-flash`	20GB+	19GB	Fast response, automation
Ministral	`ministral-3:8b`	8GB+	6GB	Lightweight, daily use
GPT-OSS	`gpt-oss-20b`	16GB+	-	Open source ecosystem exclusive

⚠️ Known Issues: qwen2.5 and qwen3 currently have compatibility issues and are temporarily not recommended!

Pull Local Models

Choose a suitable model based on your graphics card's VRAM:

Option A: GLM 4.7 Flash (Recommended)

ollama pull glm-4.7-flash

Model Size: 19GB
Suitable for: GPUs with 20GB+ VRAM

Option B: Ministral 3:8b (Lightweight)

ollama pull ministral-3:8b

Model Size: 6GB
Suitable for: GPUs with 8GB+ VRAM

Configure Cloud Models (Optional)

If you want to use SOTA models but don't have API Keys for OpenAI/Anthropic/Google Gemini:

# Login to Ollama (follow on-screen instructions to connect device)
ollama signin

# Pull Google Gemini 3 Flash cloud model
ollama pull gemini-3-flash-preview:cloud

💡 Tip: Cloud models have usage limits, do not use them too frequently. See Model Guide & Deployment for details.

3️⃣ OpenClaw Installation

Note: If you have installed Ollama v0.17.0+, it will automatically install OpenClaw for you. Please skip this step and go straight to the Initial Configuration section.

Open Command Prompt as a Regular User:

Install OpenClaw

⚠️ IMPORTANT! IMPORTANT! IMPORTANT! (Saying it three times because it's important)
You MUST use Command Prompt as a Regular User to install OpenClaw!
Installing as Administrator may cause Telegram to fail to respond properly.

Open Command Prompt as a Regular User:

curl -fsSL https://openclaw.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

This command will automatically install Node.js and npm, and enter OpenClaw's Onboarding mode (initial welcome setup screen).

Initial Configuration

1. Security Confirmation

I understand this is powerful and inherently risky. Continue?
> Yes

2. Onboarding Mode

Onboarding mode
> QuickStart (Configure details later via openclaw configure.)

3. Setup Model/Auth Provider

Choose skip for now, we will configure it later:

Model/auth provider
> Skip for now

Filter models by provider
> All providers

Default model
> Enter model manually 

# Enter the model name mentioned above, e.g., ollama/glm-4.7-flash

4. Channel Configuration (Optional)

Here you can choose Skip for now, or configure Telegram directly. Assuming we are configuring:

Select channel (QuickStart)
> ● Telegram (Bot API) (not configured)

Enter Telegram bot token
>>> 1234567890:ABCDEFGHIJKLMNOPQRSTUVWXYZ123456789

💡 How to get a Token? Please refer to Telegram Bot Setup

5. Skills Store

Configure skills now? (recommended)
> No

⚠️ Windows cannot install Brew, and the skills store requires Brew, so choose No for now.

6. Enable Hooks (if prompted)

Enable hooks?
> [+] 🚀 boot-md (Run BOOT.md on gateway startup)
> [+] 📎 bootstrap-extra-files (Inject additional workspace bootstrap files via glob/path patterns)
> [+] 📝 command-logger (Log all command events to a centralized audit file)
> [+] 💾 session-memory (Save session context to memory when /new or /reset command is issued)

Press the Spacebar to select all three items, then press Enter.

7. Record Web UI Info

After installation, it will display:

Control UI:
  Web UI: http://127.0.0.1:18789/
  Web UI (with token): http://127.0.0.1:18789/?token=xxxxxxxxxx
  Gateway WS: ws://127.0.0.1:18789
  Gateway: reachable

🔑 Important: Remember the URL with the token!

8. Install Shell (if prompted)

Enable zsh shell completion for openclaw?
> No

⚠️ Windows cannot use zsh, so choose No for now.

9. Start Gateway Service

At this time, OpenClaw will open the browser and enter the Gateway Dashboard. If the Gateway Service is not running, the content will not display properly. Please press Ctrl+C to stop the OpenClaw window first, then:

Open Command Prompt as Administrator:

openclaw gateway install

At this point, the OpenClaw Gateway Service will be installed and started, and it will run automatically upon Windows startup.

10. Configure OpenClaw to use Ollama

Ollama v0.15.3+ New Feature: You can configure OpenClaw's Ollama settings to make it use the local model.

Since the Ollama local model hasn't been applied yet, please close the Gateway window (if it appears) by pressing Ctrl+C, then enter:

ollama launch openclaw

After setting up the model, this screen will appear, answer yes to continue:

This will modify your OpenClaw configuration:
  C:\Users\<your username>\.openclaw\openclaw.json
Backups will be saved to C:\Users\<your username>\AppData\Local\Temp\ollama-backups/

Proceed? (y/n) yes

💡 Recommendation: After executing this, restart your computer first to ensure the Gateway Service boots up properly automatically.

📝 Changing models later: Please refer to the Useful Tips below.

11. Test Gateway

After restarting, verify that the Gateway Service is running (if not, just run ollama launch openclaw), then open your browser and visit:

http://127.0.0.1:18789/?token=xxxxxxxxxx

Enter any message in the Chat UI, Ollama will load the model in the background and respond.

✅ If the AI replies normally, basic setup is complete!

4️⃣ Advanced Configuration

Telegram Bot Setup

Create a Telegram Bot

Search for and add @BotFather in Telegram.
Send the command /newbot, follow prompts to set a bot name.
Example: openclaw-bot (change if taken)
BotFather will reply:

Done. Congratulations on your new bot...

Use this token to access the HTTP API:
1234567890:ABCDEFGHIJKLMNOPQRSTUVWXYZ123456789

🔑 Keep this Token safe, you'll need it for configuration later!

Pair Telegram Channel

Open the bot channel on Telegram on your phone, see if there's the following message (If not, send a random message).

OpenClaw: access not configured.

Your Telegram user id: 1234567890
Pairing code: abcdefgh

Run the pairing command on your PC:

openclaw pairing approve telegram abcdefgh

(Replace abcdefgh with your pairing code)

Send a message again to test.

✅ The Bot should reply normally now! 🎉

Other Advanced Settings (Optional)

Open Command Prompt as a regular user:

openclaw config

Gateway Location

Where will the Gateway run?
> ● Local (this machine)

Web Tools Configuration

Select sections to configure
> ● Web tools (Configure Brave search + fetch)

Enable web_search (Brave Search)?
> ○ Yes / ● No

💡 This requires a Brave API Key (can be applied separately), choose No for now.

Enable web_fetch (keyless HTTP fetch)?
> ● Yes / ○ No

🗑️ Complete Removal Guide

If you need to completely remove OpenClaw / Moltbot / Clawdbot:

Open PowerShell as Administrator

# Complete removal (including all data)
openclaw uninstall --all --yes --non-interactive
# OR
moltbot uninstall --all --yes --non-interactive
# OR
clawdbot uninstall --all --yes --non-interactive

# Remove npm packages
npm uninstall -g openclaw
# OR
npm uninstall -g moltbot
# OR
npm uninstall -g clawdbot

📄 Configuration File Reference

File Path

%USERPROFILE%\.openclaw\openclaw.json

Example Configuration

{
  "models": {
    "providers": {
      "ollama": {
        "baseUrl": "http://127.0.0.1:11434/v1",
        "apiKey": "ollama-local",
        "api": "openai-completions",
        "models": [
          {
            "id": "ollama/glm-4.7-flash",
            "name": "GLM 4.7 Flash",
            "reasoning": true,
            "input": ["text"],
            "cost": {
              "input": 0,
              "output": 0,
              "cacheRead": 0,
              "cacheWrite": 0
            },
            "contextWindow": 262000,
            "maxTokens": 16384
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "ollama/glm-4.7-flash",
        "fallbacks": ["ollama/gemini-3-flash-preview:cloud"]
      },
      "workspace": "C:\\Users\\USER\\.openclaw\\workspace",
      "compaction": {
        "mode": "safeguard"
      },
      "maxConcurrent": 4,
      "subagents": {
        "maxConcurrent": 8
      }
    }
  },
  "gateway": {
    "mode": "local",
    "auth": {
      "mode": "token",
      "token": "YOUR_TOKEN_HERE"
    },
    "port": 18789,
    "bind": "loopback"
  },
  "channels": {
    "telegram": {
      "enabled": true,
      "botToken": "YOUR_BOT_TOKEN_HERE"
    }
  },
  "hooks": {
    "internal": {
      "enabled": true,
      "entries": {
        "boot-md": { "enabled": true },
        "command-logger": { "enabled": true },
        "session-memory": { "enabled": true }
      }
    }
  }
}

🎯 Quick Reference

Command	Purpose
`ollama --version`	Check Ollama version
`ollama pull <model>`	Pull a model
`ollama launch openclaw`	Configure OpenClaw to use Ollama
`openclaw config`	Enter configuration UI
`openclaw models list`	View list of currently configured models
`openclaw gateway install`	Install Gateway service
`openclaw gateway start`	Start Gateway service
`openclaw pairing approve telegram <code>`	Pair Telegram channel
`openclaw security audit --deep`	Deep security audit
`openclaw uninstall --all`	Complete removal

💡 Useful Tips

Prevent Ollama from automatically unloading models

Add to your environment variables:

OLLAMA_KEEP_ALIVE=-1

This prevents Ollama from automatically unloading models after 5 minutes of inactivity, speeding up subsequent conversations.

Configure Ollama for parallel requests

If you need OpenClaw's advanced applications like Multi-Agents or Multi-Sessions, you need multiple concurrent LLM calls. Therefore, you must increase Ollama's parallel requests number: