在腾讯云轻量服务器部署 OpenClaw + 天翼云智谱免费GLM-5 + 飞书
前言
本文记录如何在腾讯云轻量应用服务器(2核2G)上部署 OpenClaw,接入天翼云免费 GLM-5 大模型,并通过飞书实现对话交互。整套方案成本低、配置简单,适合个人助手场景。
一、服务器准备
1.1 腾讯云配置
| 配置项 | 规格 |
|---|---|
| CPU | 2 核 |
| 内存 | 2 GiB |
| 带宽 | 1 Mbps |
| 系统 | Ubuntu 22.04 LTS |
1Mbps 带宽对于文字交互完全够用,GLM-5 模型通过 API 调用,不占用本地算力。
1.2 连接服务器
ssh ubuntu@your-server-ip
二、环境安装
2.1 安装 Node.js
OpenClaw 需要 Node.js 18+,推荐使用 nvm 管理:# 安装 nvm
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.7/install.sh | bash
source ~/.bashrc
安装 Node.js 22
nvm install 22
nvm use 22
node -v # 确认版本
2.2 通过 npm 安装 OpenClaw
npm install -g openclaw
安装完成后验证:openclaw --version
三、初始化 OpenClaw
3.1 创建工作目录
mkdir -p ~/.openclaw/workspace
cd ~/.openclaw/workspace
3.2 初始化配置
openclaw init
按提示完成基础配置。
四、配置天翼云 GLM-5 模型
4.1 获取天翼云 API 凭证
-
登录 息壤一体化智算服务平台
-
获取 baseUrl、model_id、appKey
选择模型推理服务=>服务接入=>API快捷接入=>确认创建并选择=>选择模型(智谱GLM-5)
4.2 配置模型(仅供参考)
编辑 ~/.openclaw/openclaw.json
{
"meta": {
"lastTouchedVersion": "2026.2.24",
"lastTouchedAt": "2026-02-25T16:16:36.207Z"
},
"wizard": {
"lastRunAt": "2026-02-25T16:16:36.178Z",
"lastRunVersion": "2026.2.24",
"lastRunCommand": "onboard",
"lastRunMode": "local"
},
"auth": {
"profiles": {
"ctyun:default": {
"provider": "ctyun",
"mode": "api_key"
}
}
},
"models": {
"mode": "merge",
"providers": {
"ctyun": {
"baseUrl": "https://wishub-x6.ctyun.cn/v1",
"api": "openai-completions",
"apiKey": "天翼云appKey",
"authHeader": true,
"models": [
{
"id": "天翼云model_id",
"name": "天翼云GLM-5",
"reasoning": false,
"input": [
"text",
"image"
],
"cost": {
"input": 0,
"output": 0
},
"contextWindow": 128000,
"maxTokens": 4096
}
]
}
}
},
"agents": {
"defaults": {
"model": {
"primary": "ctyun/天翼云model_id"
},
"models": {
"zai/glm-4.7": {
"alias": "GLM"
},
"ctyun/天翼云model_id": {
"alias": "天翼云GLM-5"
}
},
"workspace": "/home/ubuntu/.openclaw/workspace",
"compaction": {
"mode": "safeguard"
},
"maxConcurrent": 4,
"subagents": {
"maxConcurrent": 8
}
}
},
"messages": {
"ackReactionScope": "group-mentions"
},
"commands": {
"native": "auto",
"nativeSkills": "auto",
"restart": true,
"ownerDisplay": "raw"
},
"session": {
"dmScope": "per-channel-peer"
},
"hooks": {
"internal": {
"enabled": true,
"entries": {
"session-memory": {
"enabled": true
}
}
}
},
"channels": {
"feishu": {
"enabled": true,
"appId": "飞书appid",
"appSecret": "飞书appSecret",
"domain": "feishu",
"groupPolicy": "open"
}
},
"gateway": {
"port": 18789,
"mode": "local",
"bind": "loopback",
"auth": {
"mode": "token",
"token": "openclaw自动生成的token"
},
"tailscale": {
"mode": "off",
"resetOnExit": false
},
"nodes": {
"denyCommands": [
"camera.snap",
"camera.clip",
"screen.record",
"calendar.add",
"contacts.add",
"reminders.add"
]
},
"controlUi": {
"allowedOrigins": ["*"]
}
},
"skills": {
"install": {
"nodeManager": "npm"
}
},
"plugins": {
"entries": {
"feishu": {
"enabled": true
}
},
"installs": {
"feishu": {
"source": "npm",
"spec": "@openclaw/feishu",
"installPath": "/home/ubuntu/.openclaw/extensions/feishu",
"version": "2026.2.24",
"resolvedName": "@openclaw/feishu",
"resolvedVersion": "2026.2.24",
"resolvedSpec": "@openclaw/feishu@2026.2.24",
"integrity": "sha512-/YNWeOH56WU2WvHD+5zu4sxRzlBoK5Vwcg8eQV4hNj2xV3fDDsdV/laAWnbzjBccCgBv0zXrEpeiK/uZ/EcZ6A==",
"shasum": "a32f577fc752f9dceb6b058cc8e56e05e139c1df",
"resolvedAt": "2026-02-25T15:51:30.309Z",
"installedAt": "2026-02-25T15:51:33.559Z"
}
}
}
}
五、接入飞书
5.1 创建飞书应用
-
访问 飞书开放平台
-
创建企业自建应用
-
记录 App ID 和 App Secret
5.2 配置权限
在飞书应用后台,添加以下权限:
-
contact:user.base:readonly - 获取用户基本信息
-
im:message - 消息相关权限
-
im:message:send_as_bot - 以应用身份发消息
5.3 配置事件订阅
-
在「事件订阅」页面添加请求地址:
-
订阅事件:
- im.message.receive_v1 - 接收消息
六、启动服务
6.1 启动 Gateway
openclaw gateway start
七、验证部署
7.1 检查状态
openclaw status
7.2 飞书测试
在飞书中找到你的应用,发送消息:你好
收到 GLM-5 的回复即部署成功!
八、常见问题
8.1 Message ordering conflict - please try again. If this persists, use /new to start a fresh session.
出现这个错误排查openclaw.json文件中models是否配置正确
总结
通过腾讯云轻量服务器 + OpenClaw + 天翼云智谱免费 GLM-5 + 飞书的组合,可以低成本搭建一个功能完整的个人 AI 助手。整个部署过程约 60 分钟,适合个人开发者和小团队使用。

更多推荐

所有评论(0)