Jamdesk

Interesting product. One thing I'm struggling to understand: if tools like Claude Code can already read a codebase and generate documentation, why is automated AI documentation a separate product category? My concern is that AI-generated docs often contain mistakes, outdated assumptions, or miss important context, so many teams still want a human to review and edit the output before publishing. In that case, where do you see the biggest value coming from?

The AI-readiness bundle (llms.txt generation, MCP server, per-page Markdown export) is genuinely ahead of most docs platforms. I hand-implemented llms.txt and JSON-LD for my own site recently and the long tail of details was bigger than expected. Question: do you also handle index freshness for AI crawlers - e.g. IndexNow pings on publish - or is that left to the host?

This tool is doing great work.

I've tried a lot of documentation tools, and Jamdesk stands out for its clean approach. The Git-based workflow combined with AI chat and built-in analytics makes it much easier to maintain docs without juggling multiple services. It looks like a solid choice for teams that want fast, professional documentation sites.

This looks cool! I love the design!

The LLM integration is very smart. The source of truth is already right there, accessible to coding LLMs, so you might as well use it to automate documentation.

Huge congrats on the launch! The UI and onboarding experience look incredibly smooth from the screenshots alone. Did you use any specific framework for those micro-interactions, or is it all custom? This seems like it perfectly solves a major friction point for standard workflows. Excellent work!

this is a really interesting product

Amazing product , keep going Evan !

我感觉确实挺不错的，你们觉得怎么样？特别是UI

The positioning around documentation people actually enjoy is strong. The most compelling use case for me is reducing friction after onboarding: teams often have docs, but new users still cannot find the right answer. A short example workflow or before-after doc page would make the value even clearer.

Bundling the MCP server and llms.txt generation directly into the docs build is the standout here - most teams bolt those on manually and the per-page Markdown export detail is easy to underestimate until you need it. The OpenAPI playground with multi-language examples is what would sell me, since hand-maintaining a reference next to a spec always drifts. One question: when the OpenAPI spec updates, does the generated reference (and the MCP server's exposed tools) regenerate automatically on the next Git push, or is there a separate sync step? Drift between spec and published docs is the usual failure mode.

documentation that users actually like reading is a harder problem than it sounds. most companies end up with a support ticket problem that is really a docs findability problem. if the search is good and the content is structured around tasks rather than features, a lot of those tickets disappear. curious how you handle version control for docs when the product changes fast.

Congrats on the launch. Great product to help small SaaS startups

MDX living in your Git repo is the right call — docs next to the code actually stay current instead of rotting. Two Qs: is the AI chat grounded strictly on your own docs (and does it cite the source page it pulled from?), and how painful is migrating an existing Docusaurus/Mintlify site?

A necessary tool for current market

Clean docs platform with built-in llms.txt and MCP server. The AI-readiness angle for documentation is underrated. Most SaaS tools have docs that AI agents can't read at all. Nice launch.