Developers hate Documentation - so I'm building an LLM Agent to solve that

One of the most painful things I've dealt with (both at work and while building another startup) is maintaining accurate user facing API & SDK docs.

Auto-generators like Swagger (OpenAPI) sound great in theory… but in reality, most codebases are too messy or inconsistent for them to work reliably. I’ve lost count of the number of times the docs just didn't match what the backend actually did.

And when you turn to nicer-looking doc generators like Mintlify or Docusaurus, you're expected to edit and maintain everything manually. Which means unless you're investing time or resources, your docs go stale.

So I started working on something I’m calling DeepDocs, an LLM-powered agent that has a rich context awareness of your codebase and automatically updates your docs whenever the underlying code changes.

The core idea:

• Parses your source code
• Detects changes to request/response shapes, params, and endpoints
• Updates your user facing docs (API/SDK usage, tutorials etc.)
• Works even for codebases without OpenAPI specs

Right now it’s just a landing page and some early R&D, but I'm validating if this is a pain others feel strongly enough about.

If you've ever had to write or maintain docs, I’d love to hear your thoughts.
Would you use something like this?
What would you expect it to get right?

Open to feedback, questions, pushback—all of it.

Here’s the site if you're interested: https://deepdocs.dev