update readme and mcp demo

This commit is contained in:
Sun Tao 2025-03-18 20:06:27 +08:00
parent 8c5a5f7fa9
commit 4da8ecdb20
4 changed files with 50 additions and 31 deletions

View File

@ -422,6 +422,30 @@ Here are some tasks you can try with OWL:
OWL's MCP integration provides a standardized way for AI models to interact with various tools and data sources:
Before using MCP, you need to install Node.js first.
### **Install Node.js**
### Windows
Download the official installer: [Node.js](https://nodejs.org/en).
Check "Add to PATH" option during installation.
### Linux
```bash
sudo apt update
sudo apt install nodejs npm -y
```
### Mac
```bash
brew install node
```
### **Install Playwright MCP Service**
```bash
npm install -g @executeautomation/playwright-mcp-server
npx playwright install-deps
```
Try our comprehensive MCP example in `examples/run_mcp.py` to see these capabilities in action!
## Available Toolkits

View File

@ -410,6 +410,29 @@ OWL 将自动调用与文档相关的工具来处理文件并提取答案。
## 模型上下文协议MCP
OWL 的 MCP 集成为 AI 模型与各种工具和数据源的交互提供了标准化的方式。
在使用MCP前需要先安装Node.js。
### **安装 Node.js**
### Windows
下载官方安装包:[Node.js](https://nodejs.org/zh-cn)。
安装时,勾选 "Add to PATH" 选项。
### Linux
```bash
sudo apt update
sudo apt install nodejs npm -y
```
### Mac
```bash
brew install node
```
### **安装 MCP 服务**
```bash
npm install -g @executeautomation/playwright-mcp-server
npx playwright install-deps
```
查看我们的综合示例 `examples/run_mcp.py` 来体验这些功能!

View File

@ -1,13 +1,5 @@
{
"mcpServers": {
"desktop-commander": {
"command": "npx",
"args": [
"-y",
"@wonderwhy-er/desktop-commander",
"--force-file-protocol"
]
},
"playwright": {
"command": "npx",
"args": ["-y", "@executeautomation/playwright-mcp-server"]

View File

@ -22,27 +22,8 @@ Environment Setup:
2. MCP Server Setup:
2.1 MCP Desktop Commander (File System Service):
Prerequisites: Node.js and npm
```bash
# Install MCP service
npx -y @smithery/cli install @wonderwhy-er/desktop-commander --client claude
npx -y @wonderwhy-er/desktop-commander setup --force-file-protocol
# Configure in owl/examples/mcp_servers_config.json:
{
"desktop-commander": {
"command": "npx",
"args": [
"-y",
"@wonderwhy-er/desktop-commander",
"--force-file-protocol"
]
}
}
```
2.2 MCP Playwright Service:
2.1 MCP Playwright Service:
```bash
# Install MCP service
npm install -g @executeautomation/playwright-mcp-server
@ -59,7 +40,7 @@ Environment Setup:
}
```
2.3 MCP Fetch Service (Optional - for better retrieval):
2.2 MCP Fetch Service (Optional - for better retrieval):
```bash
# Install MCP service
pip install mcp-server-fetch
@ -168,8 +149,7 @@ async def main():
question = (
"I'd like a academic report about Andrew Ng, including "
"his research direction, published papers (At least 3),"
" institutions, etc.Then organize the report in Markdown "
"format and save it to my desktop"
" institutions, etc. "
)
# Connect to all MCP toolkits