大模型
philschmid.de ·
60秒内用Gemini设置OpenClaw 2.0
通过五个命令安装OpenClaw 2.0，连接Gemini 3.7 Flash，并在终端和网页仪表板中启用Google搜索接地。
AI推荐理由 Philipp Schmid提供OpenClaw 2.0与Gemini的快速安装指南，实用性强，但内容较浅。
70
正文
在60秒内设置带Gemini的OpenClaw 2.0
2026年8月31日4分钟阅读
五个命令即可安装OpenClaw 2.0（v2026.8.1），连接Gemini 3.7 Flash（我们用于编码和代理的最快（高达+300 token/秒）代理模型），并在不到一分钟内开始使用Google搜索接地进行聊天。
什么是OpenClaw 2.0？
OpenClaw是一个真正做事的个人AI助手。它运行在你的设备上，在你的聊天中，遵循你的规则。
随处聊天：从浏览器控制UI、Telegram、Discord、WhatsApp或Slack聊天。
集成工具：浏览器自动化、终端执行、文件系统访问和网络搜索接地。
记忆：跨会话保持记忆，因此它知道你的偏好、项目和人员。
OpenClaw 2.0（v2026.8.1）是迄今为止最大的更新：更简单的安装，重建的浏览器应用程序直接打开进入对话，以及你可以通过与你的Claw交谈来完成的设置。完整说明：v2026.8.1。
先决条件
Node.js：Node 26（推荐）或Node 22.22.3+ / 24.15+（node --version）。
Gemini API密钥：从Google AI Studio获取免费密钥。
1. 安装OpenClaw 2.0
通过npm全局安装OpenClaw：
Bash
npm install -g openclaw@2026.8.1
验证你的安装：
Bash
openclaw --version
# 输出：OpenClaw 2026.8.1 (ea80657)
2. 配置Gemini身份验证
在当前shell中导出你的Gemini API密钥：
Bash
export GEMINI_API_KEY="你的-gemini-api密钥-这里"
要将其持久化到OpenClaw的本地配置文件存储中：
Bash
openclaw onboard --non-interactive --accept-risk --skip-health \
--mode local \
--auth-choice gemini-api-key \
--gemini-api-key "$GEMINI_API_KEY"
（注意：你可以安全地忽略任何关于配置Brave API密钥的入门提示——Gemini内置的Google搜索接地会自动使用。）
3. 将Gemini 3.7 Flash设置为默认
将Gemini 3.7 Flash设置为默认代理模型：
Bash
openclaw models set google/gemini-3.7-flash
（注意：如果OpenClaw显示模型不在本地目录中的警告，可以安全忽略。该模型在下一步中会动态解析并针对Google AI Studio进行验证。）
验证模型发现和身份验证状态：
Bash
openclaw models list --provider google
4. 通过CLI测试推理
在终端中运行独立的代理回合：
Bash
openclaw agent --local --agent main --message "你好！你运行的是什么模型？"
输出：
我在google/gemini-3.7-flash上运行。
我是你在OpenClaw中的新助手！你想怎么称呼我？
5. 启动网关并启动控制UI
将网关设置为本地模式并启动它：
Bash
openclaw config set gateway.mode local
openclaw gateway run
（要作为持久后台守护进程运行，请改用：openclaw gateway install && openclaw gateway start。要停止它：openclaw gateway stop --force）
在浏览器中打开控制UI：
Bash
openclaw dashboard
浏览器将打开到http://127.0.0.1:18789/，并使用你的本地会话令牌进行身份验证：
下一步
Google搜索接地（默认启用）
当存在GEMINI_API_KEY时，Google搜索接地默认启用。无需额外配置或搜索密钥。
直接通过CLI测试：
Bash
openclaw capability web search --query "最新的gemini模型"
在聊天会话中，每当需要新鲜上下文时，Gemini会自动运行带引用的网络搜索。
参见：OpenClaw中的Google搜索接地
利用其他Gemini功能
图像生成：使用openclaw models set-image google/gemini-3.1-flash-image设置Google的图像模型。
视频生成：使用google/veo-3.1-fast-generate-preview生成4-8秒的视频。
音乐生成：通过google/lyria-3-clip-preview进行提示驱动的音乐。
文本转语音：使用自然语言风格提示的Gemini批量TTS（gemini-3.1-flash-tts-preview）。
实时语音：通过Gemini Live API（gemini-3.1-flash-live-preview）进行低延迟双向语音流。
完整的Google文档。
连接聊天渠道
使用简单的CLI命令将你的Gemini助手连接到最喜欢的消息应用程序：
浏览所有渠道：openclaw channels list --all
交互式设置：openclaw channels add
Telegram：openclaw channels add --channel telegram --token ""
WhatsApp（QR码）：openclaw channels login --channel whatsapp
Discord：openclaw channels add --channel discord --token ""
验证渠道状态：openclaw channels status --probe
完整的聊天渠道文档。
查看原文（中文正文为 AI 翻译）
Setup OpenClaw 2.0 with Gemini in Under 60 Seconds
August 31, 20264 minute read
Five commands to install OpenClaw 2.0 (v2026.8.1), connect Gemini 3.7 Flash, our fastest (up to +300 token/sec) agentic model for coding and agents, and start chatting with Google search grounding in under a minute.
What is OpenClaw 2.0?
OpenClaw is a personal AI assistant that actually does things. It runs on your devices, in your chats, with your rules.
Chat from anywhere: Chat from the browser Control UI, Telegram, Discord, WhatsApp, or Slack.
Integrated tooling: Browser automation, terminal execution, filesystem access, and web search grounding.
Remember: Keeps memory across sessions so it knows your preferences, projects, and people.
OpenClaw 2.0 (v2026.8.1) is the largest update yet: simpler install, a rebuilt browser app that opens straight into a conversation, and setup you can finish by talking to your Claw. Full notes: v2026.8.1.
Prerequisites
Node.js: Node 26 (recommended) or Node 22.22.3+ / 24.15+ (node --version).
Gemini API Key: Grab a free key from Google AI Studio.
1. Install OpenClaw 2.0
Install OpenClaw globally via npm:
Bash
npm install -g openclaw@2026.8.1
Verify your installation:
Bash
openclaw --version
# Output: OpenClaw 2026.8.1 (ea80657)
2. Configure Gemini Authentication
Export your Gemini API key in your current shell:
Bash
export GEMINI_API_KEY="your-gemini-api-key-here"
To persist it into OpenClaw's local profile store:
Bash
openclaw onboard --non-interactive --accept-risk --skip-health \
--mode local \
--auth-choice gemini-api-key \
--gemini-api-key "$GEMINI_API_KEY"
(Note: You can safely ignore any onboarding tip about configuring a Brave API key—Gemini's built-in Google Search grounding is used automatically.)
3. Set Gemini 3.7 Flash as Default
Set Gemini 3.7 Flash as the default agent model:
Bash
openclaw models set google/gemini-3.7-flash
(Note: If OpenClaw displays a warning that the model is not in the local catalog, it is safe to ignore. The model is dynamically resolved and validated against Google AI Studio in the next step.)
Verify model discovery and authentication status:
Bash
openclaw models list --provider google
4. Test Inference via CLI
Run a standalone agent turn in your terminal:
Bash
openclaw agent --local --agent main --message "Hello! What model are you running?"
Output:
I'm running on google/gemini-3.7-flash.
I'm your new assistant here in OpenClaw! What would you like to call me?
5. Start the Gateway & Launch Control UI
Set the gateway to local mode and start it:
Bash
openclaw config set gateway.mode local
openclaw gateway run
(To run as a persistent background daemon instead: openclaw gateway install && openclaw gateway start. To stop it: openclaw gateway stop --force)
Open the Control UI in your browser:
Bash
openclaw dashboard
The browser will open to http://127.0.0.1:18789/ authenticated with your local session token:
Next Steps
Google Search Grounding (Enabled by Default)
Google Search grounding is enabled by default when GEMINI_API_KEY is present. No extra configuration or search keys are required.
Test it directly via CLI:
Bash
openclaw capability web search --query "Latest gemini models"
In chat sessions, Gemini autonomously runs web searches with citations whenever fresh context is needed.
See: Google Search Grounding in OpenClaw
Leveraging Other Gemini Capabilities
Image Generation: Set Google's image model using openclaw models set-image google/gemini-3.1-flash-image.
Video Generation: Generate 4–8s videos with google/veo-3.1-fast-generate-preview.
Music Generation: Prompt-driven music via google/lyria-3-clip-preview.
Text-to-Speech: Gemini batch TTS with natural language style prompting (gemini-3.1-flash-tts-preview).
Realtime Voice: Low-latency bidirectional voice streams via Gemini Live API (gemini-3.1-flash-live-preview).
Full Google Documentation.
Connecting Chat Channels
Connect your Gemini assistant to your favorite messaging apps using simple CLI commands:
Browse all channels: openclaw channels list --all
Interactive setup: openclaw channels add
Telegram: openclaw channels add --channel telegram --token ""
WhatsApp (QR code): openclaw channels login --channel whatsapp
Discord: openclaw channels add --channel discord --token ""
Verify channel status: openclaw channels status --probe
Full Chat Channels Documentation.