Content
> **⚠️ Service Disruption Notice**
> The NetworkLytics server is currently in an operational disruption state.
> You can use it in a local environment by changing `NETWORKLYTICS_API_URL` to a local server address.
---
# NetworkLytics MCP
[NetworkLytics](https://networklytics.net) MCP (Model Context Protocol) server.
Claude Desktop, Cursor, and other MCP-supported AI tools allow direct access to YouTube comment network analysis results.
## Installation
```bash
pip install networklytics-mcp
```
## Claude Desktop Configuration
Add to `~/.claude/claude_desktop_config.json` or `%APPDATA%\Claude\claude_desktop_config.json`:
```json
{
"mcpServers": {
"networklytics": {
"command": "networklytics-mcp",
"env": {
"NETWORKLYTICS_API_URL": "https://networklytics.net",
"NETWORKLYTICS_API_KEY": "nly_your_api_key_here"
}
}
}
}
```
The API key can be issued on the NetworkLytics account settings page.
## Usage Examples
You can request Claude as follows:
- "Show me the analysis results for this shared link: https://networklytics.net/shared/abc-123-..."
- "Find the top 3 influencers from the analysis results"
- "What is the sentiment trend of this YouTube channel's comment community?"
## Provided Tools
| Tool | Description | Authentication |
|------|-------------|---------------|
| `get_shared_analysis` | Retrieve analysis results by shared link token | Not required |
| `get_analysis_by_id` | Retrieve results by analysis ID | API key required |
| `get_api_info` | API information and endpoint list | Not required |
## Returned Data Structure
```json
{
"video": { "title": "...", "channel": "...", "view_count": 123456 },
"network": {
"total_nodes": 1500,
"total_edges": 3200,
"density": 0.003,
"community_count": 7
},
"sentiment": {
"positive_ratio": 0.62,
"negative_ratio": 0.15,
"overall": "positive"
},
"top_influencers": [
{ "author": "username", "degree_centrality": 0.12, "comment_count": 45 }
],
"topic_keywords": ["keyword1", "keyword2"],
"ai_insights": { ... }
}
```
