OpenClaw使用体验
作者：数据人阿多 日期：2026年3月13日
小编体验结论
- 配置有点复杂
- 太消耗 tokens
- 不同模型的使用效果有差异
下面分别就以上几点体验进行详细阐述，网上现在已经有大量的安装教程，小编这里不再赘述
小编环境
- Windows11
- WSL2 Ubuntu-24.04
- OpenClaw 官方安装，安装文档：https://docs.openclaw.ai/zh-CN/install
配置有点复杂
OpenClaw 大家可以把它当做一个 软件 或者 智能系统 使用，遇到什么问题，那大概率就是配置有问题
国内用户习惯界面UI开关按钮，进行软件控制，但OpenClaw有的配置界面配置太麻烦
配置可以从多个途径进行修改：
-
提供详细的提示词，让智能体自己去修改-----容易自己把自己修改废了
-
在 WebUI中修改配置，可以理解为管理后台-----在保存配置时，OpenClaw会备份之前的配置
-
命令行进行修改，命令有点复杂
-
直接修改配置文件-----修改完需要重启网关
小编这里是WSL2中安装的OpenClaw，对应的文件是
\\wsl.localhost\Ubuntu-24.04\home\datashare\.openclaw\openclaw.json
小编的一些配置，供大家参考：
模型配置
其中 DeepSeek 的模型，OpenClaw配置模型选项中没有，所以需要自己配置
小编体验了不同平台的模型，所以列出的模型配置有点多
"models": {
"mode": "merge",
"providers": {
"deepseek": {
"baseUrl": "https://api.deepseek.com/v1",
"apiKey": "xxx",
"api": "openai-completions",
"models": [
{
"id": "deepseek-chat",
"name": "DeepSeek Chat (V3)"
},
{
"id": "deepseek-reasoner",
"name": "DeepSeek Reasoner (R1)"
}
]
},
"bailian": {
"baseUrl": "https://dashscope.aliyuncs.com/compatible-mode/v1",
"apiKey": "xxx",
"api": "openai-completions",
"models": [
{
"id": "qwen3.5-plus",
"name": "qwen3.5-plus",
"api": "openai-completions",
"reasoning": false,
"input": [
"text",
"image"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 1000000,
"maxTokens": 65536
},
{
"id": "qwen3-max-2026-01-23",
"name": "qwen3-max-2026-01-23",
"api": "openai-completions",
"reasoning": false,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 262144,
"maxTokens": 65536
},
{
"id": "qwen3-coder-next",
"name": "qwen3-coder-next",
"api": "openai-completions",
"reasoning": false,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 262144,
"maxTokens": 65536
},
{
"id": "qwen3-coder-plus",
"name": "qwen3-coder-plus",
"api": "openai-completions",
"reasoning": false,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 1000000,
"maxTokens": 65536
},
{
"id": "MiniMax-M2.5",
"name": "MiniMax-M2.5",
"api": "openai-completions",
"reasoning": false,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 196608,
"maxTokens": 32768
},
{
"id": "glm-5",
"name": "glm-5",
"api": "openai-completions",
"reasoning": false,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 202752,
"maxTokens": 16384
},
{
"id": "glm-4.7",
"name": "glm-4.7",
"api": "openai-completions",
"reasoning": false,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 202752,
"maxTokens": 16384
},
{
"id": "kimi-k2.5",
"name": "kimi-k2.5",
"api": "openai-completions",
"reasoning": false,
"input": [
"text",
"image"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 262144,
"maxTokens": 32768
}
]
},
"zai": {
"baseUrl": "https://open.bigmodel.cn/api/paas/v4",
"api": "openai-completions",
"models": [
{
"id": "glm-5",
"name": "GLM-5",
"reasoning": true,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 204800,
"maxTokens": 131072
},
{
"id": "glm-4.7",
"name": "GLM-4.7",
"reasoning": true,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 204800,
"maxTokens": 131072
},
{
"id": "glm-4.7-flash",
"name": "GLM-4.7 Flash",
"reasoning": true,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 204800,
"maxTokens": 131072
},
{
"id": "glm-4.7-flashx",
"name": "GLM-4.7 FlashX",
"reasoning": true,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 204800,
"maxTokens": 131072
}
]
}
}
}
多智能体配置
默认安装完后是一个 main 智能体，然后小编这里又添加了一个 stock 智能体，这两个智能体是分开的，后期可以各自负责处理不同 主题/领域 的内容，类似垂类的概念
"agents": {
"defaults": {
"model": {
"primary": "deepseek/deepseek-chat",
"fallbacks": [
"bailian/glm-5",
"bailian/glm-4.7",
"bailian/kimi-k2.5",
"deepseek/deepseek-reasoner",
"qwen-portal/coder-model",
"qwen-portal/vision-model",
"bailian/qwen3.5-flash",
"bailian/qwen3.5-plus",
"bailian/qwen3-max-2026-01-23",
"bailian/qwen3-coder-next",
"bailian/qwen3-coder-plus",
"bailian/MiniMax-M2.5"
]
},
"models": {
"deepseek/deepseek-chat": {
"alias": "deepseek-v3"
},
"deepseek/deepseek-reasoner": {
"alias": "deepseek-r1"
},
"qwen-portal/coder-model": {
"alias": "qwen"
},
"qwen-portal/vision-model": {},
"bailian/qwen3.5-flash": {},
"bailian/qwen3.5-plus": {},
"bailian/qwen3-max-2026-01-23": {},
"bailian/qwen3-coder-next": {},
"bailian/qwen3-coder-plus": {},
"bailian/MiniMax-M2.5": {},
"bailian/glm-5": {},
"bailian/glm-4.7": {},
"bailian/kimi-k2.5": {},
"zai/glm-4.7": {
"alias": "GLM-4.7"
},
"zai/glm-5": {
"alias": "GLM"
}
},
"workspace": "/home/datashare/.openclaw/workspace",
"compaction": {
"mode": "safeguard",
"reserveTokensFloor": 20000,
"memoryFlush": {
"enabled": true,
"softThresholdTokens": 4000,
"prompt": "Write any lasting notes to memory/YYYY-MM-DD.md; reply with NO_REPLY if nothing to store.",
"systemPrompt": "Session nearing compaction. Store durable memories now."
}
},
"maxConcurrent": 4,
"subagents": {
"maxConcurrent": 8
}
},
"list": [
{
"id": "main"
},
{
"id": "stock",
"name": "stock",
"workspace": "/home/datashare/.openclaw/workspace-stock",
"agentDir": "/home/datashare/.openclaw/agents/stock/agent",
"model": "deepseek/deepseek-chat"
}
]
}
消息渠道选择的是 qqbot
qqbot 注册地址：https://q.qq.com/qqbot/openclaw/login.html
"channels": {
"qqbot": {
"enabled": true,
"accounts": {
"bot1": {
"enabled": true,
"allowFrom": [
"*"
],
"appId": "xxx",
"clientSecret": "xxx"
},
"bot2": {
"enabled": true,
"allowFrom": [
"*"
],
"appId": "xxx",
"clientSecret": "xxx"
}
}
}
}
多智能体路由配置
main 智能体通过 bot1 进行通讯，stock 智能体通过 bot2 进行通讯
"bindings": [
{
"agentId": "main",
"match": {
"channel": "qqbot",
"accountId": "bot1"
}
},
{
"agentId": "stock",
"match": {
"channel": "qqbot",
"accountId": "bot2"
}
}
]
定时任务配置
定时任务配置文件：\\wsl.localhost\Ubuntu-24.04\home\datashare\.openclaw\cron\jobs.json
其中 delivery.to 的值，qqbot 有点坑，这里的值不是qqbot 的 appId ，而是在管理后台（http://127.0.0.1:18789/sessions）会话中 Key 字段下面的小字 qqbot:c2c:xxx 中的 xxx 部分，这样定时任务执行完后，可以把信息发送到 qqbot
{
"version": 1,
"jobs": [
{
"id": "d38bd60a-8b9e-467e-a3b3-5d782ccfdb93",
"agentId": "main",
"name": "test",
"enabled": true,
"deleteAfterRun": false,
"createdAtMs": 1773303832101,
"updatedAtMs": 1773306448965,
"schedule": {
"kind": "every",
"everyMs": 300000,
"anchorMs": 1773306272809
},
"sessionTarget": "isolated",
"wakeMode": "now",
"payload": {
"kind": "agentTurn",
"message": "总结今日主要的新闻，分为国内、国外主题",
"model": "deepseek-chat",
"thinking": "high"
},
"delivery": {
"mode": "announce",
"channel": "qqbot",
"to": "xxx",
"accountId": "bot1",
"bestEffort": true
},
"failureAlert": false,
"state": {
"lastRunAtMs": 1773306329473,
"lastRunStatus": "ok",
"lastStatus": "ok",
"lastDurationMs": 54914,
"lastDelivered": true,
"lastDeliveryStatus": "delivered",
"consecutiveErrors": 0
}
}
]
}
太消耗 tokens
小编是在 deepseek 充值了 10元 ，刚开始使用了一会，后来切换为智普的模型，因注册智普的账号，送了一些模型的tokens，其中700万 tokens 用了不到1天，提示没有了
小编创建的2个智能体，主要使用内容如下：
main 智能体
- 学习了小编的github博客：https://datashare-duo.github.io/datashare
- 让智能体把学习到的内容保存到记忆Memory里面，学习到的技能也保存到SKILL里面
- 考试智能体学习到的关于 pandas 的知识点
stock 智能体
- 分析了中国电信、北京文化，2个股票的技术走势，何时进行调整
- 创建定时任务用于实时跟进盘中走势
不同模型的使用效果有差异
因智普送的tokens，是低阶模型GLM-4.7，回答的股票分析数据有明显错误：
- 中国电信，2026-03-12的收盘价格不是6.05元，而是6.00元
- 北京文化，2026-03-12的收盘价格不是4.15元，而是4.17元
后来把模型切换为 DeepSeek ，再次让大模型回答，数据正确的
历史相关文章
以上是自己实践中遇到的一些问题，分享出来供大家参考学习，欢迎关注微信公众号：DataShare ，不定期分享干货
