🌽 Personal MCP Intelligence Platform v2.0

A high-performance Model Context Protocol (MCP) server integrated with a Django-powered analytics dashboard. Features the Corn Intelligence toolset for comprehensive AI interaction tracking, AST code analysis, and real-time token efficiency metrics.

🚀 Key Features

1. Intelligence Dashboard

• Real-time Monitoring: Every tool call is logged with latency, token usage, and efficiency metrics.

• Auto-Reload: Dashboard refreshes every 15 seconds with a visual countdown indicator.

• Token Analytics: Visualize direct token costs vs. context savings with efficiency scoring.

• Session Explorer: Browse and drill into past sessions at /sessions/.

2. Corn Intelligence Toolset (v2.0) — 19 Tools

3. AST Code Intelligence Engine

• Auto-Indexing: corn_list_repos automatically scans and indexes Python source files on first call.

• Symbol Graph: Stores Function, Class, and Method nodes with CALLS / INHERITS relationships in MySQL.

• Blast Radius: corn_code_impact identifies all callers transitively affected by changes to a file.

4. Token Efficiency Metrics

• Accurate Counting: Heuristic blends character density and word count — max(chars/4, words × 1.35) — for reliable estimation across code and natural language.

• Context Saved: For search tools (memory_search, knowledge_search, code_search), tokens saved = total DB tokens − tokens returned. Reflects real context reduction.

• Efficiency (%): Saved / (Used + Saved). Aim for >80%!

Related MCP server: MCP of MCPs

🏛️ Architecture

IDE (Antigravity / Claude)
    │  stdio (MCP protocol)
    ▼
Docker: personal-mcp-web
    ├── Django MCP Server (run_mcp management command)
    │       └── 19 Corn Intelligence Tools
    ├── AST Indexer (mcp_server/utils/indexer.py)
    └── Django Dashboard (http://localhost:8000)
            └── Real-time Activity Log, Sessions, Token Analytics
    │
    ▼
Docker: mcp_mysql_db (MySQL 8)
    ├── Session, ToolLog
    ├── Memory, Knowledge
    └── Repository, Symbol, SymbolRelation, PlanQuality

🏃 Getting Started

1. Launch Services:

   docker-compose up -d

2. Connect IDE — add to your MCP config (mcp_config.json):

   {
     "mcpServers": {
       "personal-mcp-v2": {
         "command": "/usr/local/bin/docker",
         "args": ["exec", "-i", "personal-mcp-web", "python", "manage.py", "run_mcp", "--verbosity", "0", "--no-color", "--skip-checks"]
       }
     }
   }

3. Run Migrations (first-time setup):

   docker exec personal-mcp-web python manage.py migrate

4. Explore Dashboard: http://localhost:8000/

🛠️ Development

Apply new migrations after model changes:

docker exec personal-mcp-web python manage.py makemigrations mcp_server
docker exec personal-mcp-web python manage.py migrate mcp_server

Check Django config inside container:

docker exec personal-mcp-web python manage.py check

View live logs:

docker logs -f personal-mcp-web

Powered by Antigravity — Premium AI Engineering.