Install OpenLM AI module management on Windows

1️⃣ Install Python

  • OpenLM is Python-based. Make sure you have Python 3.9+ installed.

  • Check installation:

    python --version
    pip --version

  • If missing, install from python.org.

  • Optional: Add Python to PATH during installation.


2️⃣ Install OpenLM Module Manager

Open a Windows Command Prompt or PowerShell:

复制代码
pip install openlm

This installs the CLI and API server tools.


3️⃣ Initialize OpenLM Repository

Create the module storage folder:

复制代码
openlm init
  • This will create folders like:

    C:\Users<User>.openlm
    modules
    configs
    logs\

  • All AI modules you register will be stored here.


4️⃣ Register / Add AI Modules

Suppose you have an AI module or script (e.g., weather-agent):

复制代码
openlm add-module --name weather-agent --type agent --path "D:\project\weather-agent" --gpu 0
  • --name → module ID

  • --type → agent / model / tool

  • --path → folder with code or checkpoint

  • --gpu → GPU index if available

You can repeat this for all your AI modules.


5️⃣ List Registered Modules

复制代码
openlm list

Example output:

复制代码
ID      Name            Type       Status
1       weather-agent   agent      ready
2       code-reviewer   model      ready

6️⃣ Start the Module Server

Open a terminal in Windows 11:

复制代码
openlm serve
  • Starts a local REST API server , typically at http://localhost:8000.

  • Each module is exposed via endpoints like /run?module=weather-agent.


7️⃣ Use a Module via Python

Example: call a module to check last 7 days' weather:

复制代码
import requests

resp = requests.post("http://localhost:8000/run", json={
    "module": "weather-agent",
    "input": "Check last 7 days rain in New York"
})

print(resp.json())
  • Works for agents, models, and tools

  • Responses are JSON objects with module output.


8️⃣ Update or Remove Modules

复制代码
openlm update-module --name weather-agent --path "D:\project\weather-agent-new"
openlm remove-module --name old-module
  • Keeps your modules versioned and organized.

9️⃣ Notes for Windows Users

  • GPU usage: Make sure CUDA drivers are installed and PyTorch detects the GPU.

  • Firewall: Allow Python/Server through firewall if you want remote access.

  • Paths : Use double quotes "D:\folder\path" in CMD / PowerShell.

  • PowerShell vs CMD: Both work; PowerShell allows better scripting.


If you want, I can also make a diagram showing OpenLM on Windows + OpenClaw + GPU + modules, so you can see visually how everything interacts.

Do you want me to do that?

相关推荐
AI算力2 小时前
让安全驾驶有“AI”相伴|腾视科技DMS视频监控一体机,守护每一次出行
人工智能·科技·ai·车载终端·ai算力·车载智能终端·车载dvr
泯泷2 小时前
超越基准:Claude Opus 4.5 的 4 个惊人洞察
人工智能
果果软件2 小时前
2026 后台管理系统选型:若依、JeecgBoot、芋道、NocoBase 和 GewuMind Boot 怎么选?
java·人工智能·后端·ai·开源
Csvn2 小时前
刚开始用 WorkBuddy,第一步该做什么?
人工智能
一水鉴天3 小时前
同一与唯一:AppBus 外壳与 DevOps 内核 之1 20261121(元宝)
人工智能·机器学习
Wellijohn3 小时前
一个 Android 老兵的十年开源路:从 167 star 的滑动控件,到本地优先的 AI Agent
android·人工智能·开源
ZONGYUAN-ROOT3 小时前
【API配额】API 配额监控告警不生效?用 Python 监控读数变化量,抓住静态阈值漏掉的「假重置」
人工智能
Wellijohn4 小时前
我把公司内部的 12 个 CRUD 接口交给了 AI Agent,然后省出了两周
人工智能·python·mcp
伞伞悦读5 小时前
【第92期】逻辑回归的分数概率与阈值的实验与排查
人工智能·python·深度学习·机器学习