MCP (Model Context Protocol)
The Model Context Protocol (MCP) revolutionizes how Pochi interacts with external services, databases, and APIs. Think of MCP as a universal adapter that connects Pochi to virtually any data source or tool, dramatically expanding what your AI assistant can accomplish.
What is MCP?
Model Context Protocol is an open standard that allows AI applications to securely connect with external resources. It's like having a universal translator that lets Pochi communicate with databases, APIs, cloud services, and custom tools using a common language.
Key Concepts
- MCP Servers: Small programs that expose specific functionality to Pochi
- Tools: Functions that Pochi can execute through MCP servers
- Resources: Read-only data sources like files or database queries
- Transport: Communication method (local processes or remote HTTP endpoints)
How It Works
- You configure MCP servers in Pochi's settings
- Pochi discovers available tools from each connected server
- When you chat with Pochi, it can use these tools to accomplish tasks
- All tool executions require your explicit approval for security
Real-World User Stories
🔍 Automated PR Review Workflow
"Streamline code review process with intelligent GitHub integration"
Setup: Connect GitHub MCP server with repository access
Result: Pochi can automatically review PRs, check code quality, suggest improvements, and manage PR lifecycle Value: Faster code review cycles with consistent quality standards
🎭 Component Validation with Playwright
"Automatically test and validate UI components across different scenarios"
Setup: Connect Playwright MCP server for automated browser testing
Result: Pochi generates components and automatically validates styling, interactions, and responsive behavior through Playwright tests Value: Consistent component quality with automated visual regression testing
📊 Database-Driven Development
"Query, analyze, and deploy with integrated database operations"
Setup: Connect PostgreSQL and AWS MCP servers Result: Pochi can query your database, analyze performance, and deploy directly to AWS Value: Streamlined development workflow with AI assistance
Step-by-Step Setup Guide
Manual VSCode Settings Configuration
For advanced users who prefer direct configuration:
Step 1: Open VSCode Settings
- Open VS Code
- Press
Cmd/Ctrl + ,
to open Settings - Search for "pochi mcp" or navigate to Extensions → Pochi
Step 2: Edit settings.json
Add to your VS Code settings.json
:
{
"pochi.mcpServers": {
"playwright": {
"command": "npx",
"args": ["playwright-mcp-server"]
},
"context7": {
"command": "npx",
"args": ["@upstash/context7-mcp"]
},
"github": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-github"],
"env": {
"GITHUB_PERSONAL_ACCESS_TOKEN": "your-github-token"
}
}
}
}
Step 3: Restart VSCode
Restart VSCode for the changes to take effect.
Configuration Examples
Local MCP Servers (Stdio Transport)
Filesystem Access
{
"filesystem": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-filesystem",
"/path/to/allowed/files"
],
"env": {
"HOME": "/Users/yourname"
}
}
}
Database Connection
{
"postgres": {
"command": "python",
"args": ["-m", "mcp_postgres", "--connection-string"],
"env": {
"DATABASE_URL": "postgresql://user:pass@localhost:5432/db"
}
}
}
Web Search
{
"brave-search": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-brave-search"],
"env": {
"BRAVE_API_KEY": "your-brave-api-key"
}
}
}
Remote MCP Servers (HTTP Transport)
Custom API Server
{
"my-api": {
"url": "https://api.mycompany.com/mcp",
"headers": {
"Authorization": "Bearer your-api-token",
"X-API-Version": "v1"
}
}
}
SSE (Server-Sent Events) Server
{
"realtime-data": {
"url": "https://data.example.com/mcp/sse",
"headers": {
"Authorization": "Bearer your-token"
}
}
}
Advanced Configuration
Server Control Options
{
"my-server": {
"command": "node",
"args": ["server.js"],
"disabled": false, // Enable/disable entire server
"disabledTools": ["dangerous_tool"], // Disable specific tools
"env": {
"NODE_ENV": "production"
}
}
}