django_mcp — MCP & Admin Chat Agent

django_mcp does two independent things that share a configuration model and a tool registry:

1. MCP protocol server — a JSON-RPC 2.0 endpoint that lets external AI clients (Claude Desktop, Cursor, custom agents) read and query your Django application’s data.
2. Admin chat agent — an LLM-powered chat interface embedded directly in the Django admin dashboard. Superusers can ask natural-language questions and get answers backed by real Django queries.

Both subsystems use the same global tool registry, so any tool you register is available in both contexts immediately.

What you can do with it

• Connect Claude Desktop, Cursor, or any MCP-compatible client to your Django app and let it query models, run management commands, and inspect the app structure.
• Let admin users ask “How many users signed up this week?” or “Show me the 10 largest orders” directly from the admin panel — no SQL, no custom views.
• Build custom tools that wrap any Django logic and expose them to both external MCP clients and the admin chat in one registration step.
• Track every tool call in MCPAuditLog for compliance and debugging.

Quick start

# api/config.py
from django_cfg import DjangoConfig
from django_cfg.modules.django_mcp.__cfg__ import DjangoMCPModuleConfig
 
class MyConfig(DjangoConfig):
    mcp = DjangoMCPModuleConfig(
        enabled=True,
        access_key="your-secret-key",
    )

That’s enough to start the MCP server at POST /cfg/mcp/ and enable the admin chat. Add exposed_apps to let agents query your models.