# π HTTP Server & Web Testing Guide
## Overview
FastMCP v2 NotebookLM server cΓ³ thα» chαΊ‘y vα»i **HTTP transport** thay vΓ¬ STDIO, cho phΓ©p:
- π **Web-based testing** của MCP tools
- π **Multiple client connections**
- π **HTTP API access** Δα» integrate vα»i web apps
- π§ͺ **Easy debugging** qua browser/Postman
## π CΓ‘ch chαΊ‘y HTTP Server
### 1. STDIO Mode (Default)
```bash
# Traditional STDIO transport
notebooklm-mcp --config notebooklm-config.json server --fastmcp
```
### 2. HTTP Mode
```bash
# HTTP transport on port 8001
notebooklm-mcp --config notebooklm-config.json server --fastmcp --transport http --port 8001 --headless
```
### 3. SSE Mode
```bash
# Server-Sent Events transport
notebooklm-mcp --config notebooklm-config.json server --fastmcp --transport sse --port 8002 --headless
```
## π Endpoints & Access
### HTTP Server URLs
| Transport | URL | Description |
|-----------|-----|-------------|
| **HTTP** | `http://localhost:8001/mcp` | MCP protocol endpoint |
| **SSE** | `http://localhost:8002/` | Server-Sent Events |
| **Custom** | `http://localhost:PORT/health` | Health check (if implemented) |
### MCP Protocol via HTTP
**URL**: `http://localhost:8001/mcp`
**Method**: `POST`
**Headers**: `Content-Type: application/json`
## π Available MCP Tools
| Tool | Description | Parameters |
|------|-------------|------------|
| `healthcheck` | Server health status | None |
| `send_chat_message` | Send message to NotebookLM | `message: str`, `wait_for_response: bool` |
| `get_chat_response` | Get streaming response | `timeout: int` |
| `chat_with_notebook` | Complete chat interaction | `message: str`, `notebook_id?: str` |
| `navigate_to_notebook` | Switch notebooks | `notebook_id: str` |
| `get_default_notebook` | Get current notebook | None |
| `set_default_notebook` | Set default notebook | `notebook_id: str` |
| `get_quick_response` | Get current response | None |
## π§ͺ Testing vα»i HTTP Clients
### 1. List Available Tools
```bash
curl -X POST http://localhost:8001/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/list",
"params": {}
}'
```
### 2. Call Healthcheck Tool
```bash
curl -X POST http://localhost:8001/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "healthcheck",
"arguments": {}
}
}'
```
### 3. Send Chat Message
```bash
curl -X POST http://localhost:8001/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 3,
"method": "tools/call",
"params": {
"name": "send_chat_message",
"arguments": {
"message": "Hello from HTTP client!",
"wait_for_response": false
}
}
}'
```
## π Python Client Example
```python
import httpx
import asyncio
async def test_mcp_http():
async with httpx.AsyncClient() as client:
# List tools
response = await client.post(
"http://localhost:8001/mcp",
json={
"jsonrpc": "2.0",
"id": 1,
"method": "tools/list",
"params": {}
}
)
tools = response.json()
print(f"Available tools: {len(tools['result']['tools'])}")
# Call healthcheck
response = await client.post(
"http://localhost:8001/mcp",
json={
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "healthcheck",
"arguments": {}
}
}
)
health = response.json()
print(f"Health status: {health['result']}")
# Run test
asyncio.run(test_mcp_http())
```
## π Web Interface Testing
### Simple HTML Test Page
```html
<!DOCTYPE html>
<html>
<head>
<title>MCP Tool Tester</title>
</head>
<body>
<h1>NotebookLM MCP Tools</h1>
<button onclick="testHealthcheck()">Test Healthcheck</button>
<div id="result"></div>
<script>
async function testHealthcheck() {
try {
const response = await fetch('http://localhost:8001/mcp', {
method: 'POST',
headers: {'Content-Type': 'application/json'},
body: JSON.stringify({
jsonrpc: "2.0",
id: 1,
method: "tools/call",
params: {
name: "healthcheck",
arguments: {}
}
})
});
const result = await response.json();
document.getElementById('result').innerHTML =
'<pre>' + JSON.stringify(result, null, 2) + '</pre>';
} catch (error) {
document.getElementById('result').innerHTML =
'<p style="color: red;">Error: ' + error.message + '</p>';
}
}
</script>
</body>
</html>
```
## π§ Client Integration Examples
### LangGraph HTTP Client
```python
from langchain_mcp_adapters.client import HttpMCPClient
client = HttpMCPClient(
url="http://localhost:8001/mcp",
transport="http"
)
tools = await client.get_tools()
model_with_tools = model.bind_tools(tools)
```
### CrewAI HTTP Integration
```python
from crewai_tools import BaseTool
import httpx
class NotebookLMHttpTool(BaseTool):
name = "notebooklm_http"
description = "Chat with NotebookLM via HTTP"
async def _arun(self, message: str):
async with httpx.AsyncClient() as client:
response = await client.post(
"http://localhost:8001/mcp",
json={
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "chat_with_notebook",
"arguments": {"message": message}
}
}
)
result = response.json()
return result["result"]["response"]
```
## π Monitoring & Debugging
### Server Logs
```bash
# Start with debug logging
notebooklm-mcp --config notebooklm-config.json server --fastmcp --transport http --debug
```
### Health Check Endpoint
```bash
# Quick health check
curl http://localhost:8001/health
```
### Tool Discovery
```bash
# List all available tools
curl -X POST http://localhost:8001/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {}}'
```
## π― Use Cases
1. **π§ͺ Development & Testing**: Easy tool testing via browser/Postman
2. **π Web Integration**: Embed NotebookLM in web applications
3. **π± Mobile Apps**: HTTP API for mobile client integration
4. **π€ Multi-Client**: Multiple AI agents connecting simultaneously
5. **π Monitoring**: Health checks and status monitoring
6. **π Load Balancing**: Scale across multiple server instances
## π¨ Production Considerations
- **π Authentication**: Add proper auth for production use
- **π‘οΈ CORS**: Configure CORS for web client access
- **π Scaling**: Use load balancer for multiple instances
- **π Monitoring**: Set up health checks and metrics
- **π HTTPS**: Use TLS certificates for secure connections
## π Summary
HTTP transport enables:
- β
**Easy testing** vα»i web tools
- β
**Multiple clients** connecting simultaneously
- β
**Web integration** for browser-based apps
- β
**API access** for any HTTP client
- β
**Better debugging** vΓ monitoring capabilities
Perfect for development, testing, vΓ production web deployments! π
