Official MCP Server for Django's documentation

Django Internals Documentation
1

Problem

  • Exposing Django’s official documentation with an MCP server that can be deployed locally or can be used from a remote dedicated URL like mcp.djangoproject.com can help the LLM or agents that developers are using widely write better Django code.
  • Right now the Django development by these AI tools are completely based on LLM general knowledge is a little outdated and does not have realtime access to Django’s documentation

Proposal

  • Exposes Django documentation as tools to AI assistants.
  • Dedicated hosted server at mcp.djangoproject.com and a local deployable version.
  • Have tools like: Search docs based on a task, Fetch version specific API Documentation,
  • Current security and performance best practices
  • Auto complete/ Edit code tool to modify/complete existing code or provide suggestions.

Existing Precedents

MCP servers are straightforward to build and maintain, they can be auto generated from the latest documentation in a periodic releases.

2 Likes
2

i said my point in discord, to mirror it:

I’m against turning django docs into a “ai friendly first” doc, if there’s going to be an ai first doc it should be held separately

documentation should be for humans first

4 Likes
3

This is a clear proposal for creating a new MCP server and tools that will make the django documentation more accessible to ai agents no where it says we are going to convert django documentation into something else.

4

to be clear I’m not disagreeing with the proposal
just pointing out that there are things we should be mindful of

5

I think it would be worth working on. I think if you want to see it landed, at this moment the best course I can think of is probably to try to implement something yourself, and if it works well we can think about the process of making it “official” and putting it in a more permanent location. Probably that needs the SC and / or ops team to chime in. But until then I think there’s a lot of value in experimenting.