LM-Kit OneThe Private AI Application Server
Promoted
Ship beautiful, fast, AI-ready documentation from plain markdown. Blume gives you search, theming, SEO, and one-command deploys on Astro and Vite.
What's great
Set it up on a small internal docs site over the weekend to replace a Docusaurus install that had gotten slow to build. The one-command deploy on Astro actually works as advertised, and the search felt instant compared to what I had before. Theming was straightforward without fighting the config. Still poking at whether it handles a really large doc set (500+ pages) as well as it handles the smaller one I tried, so holding back a star until I've pushed it further.
What needs improvement
Would like a way to preview the AI-generated summaries/llms.txt output before it goes live, rather than trusting it blindly - right now it's a bit of a black box what exactly gets exposed to crawlers.
vs Alternatives
DocusaurusDocusaurus had done the job for years but the build times had crept up as the doc set grew, and I never liked fighting the swizzle/theming system to customize anything. Blume's config felt like less to fight against for a smaller project like this one.
Ratings
Ease of use
Reliability
Value for money
Customization
Report
17 views
Mastra
the maker of @Ultracite keeps cooking.
@haydenbleasel just introduced @Blume, an open-source alternative to @Mintlify and @Fumadocs that is fast, AI-ready, and zero-config.
view source code here
Huge congratulations on the release! The decision to ship AI-ready endpoints alongside static Astro performance is a massive win for modern docs.
Does Blume support custom MDX components out of the box, or do those require extending the config? Upvoted and wishing you a great launch day!
Ultracite
@vahid_davoudi Yep custom MDX is definitely supported: https://useblume.dev/docs/configuration/customization
Dial
curious what "AI-ready" means concretely here - is it generating something like an llms.txt under the hood, or is it more about the markdown being cleanly chunked/structured for retrieval. the docs tools I've seen that use this label usually just mean well-formed markdown, but the actual pain point for me has always been staleness - docs drifting from the real API surface over time. does blume do anything to flag that drift, or is it purely a generation/theming layer on top of whatever markdown you feed it
Ultracite
@galdayan Means lots of things - llms.txt, raw markdown via negotiation headers, mcp servers, agent readability, skills. All in the pursuit of making your docs easier to read by agents. Check it out here: https://useblume.dev/docs/configuration/ai
Ultracite
@galdayan As for staleness, we have a few things to help here e.g. translations but hoping we can work on bridging the gap between product and docs more going forward, perhaps with skills.
Dial
@haydenbleasel that's a fair scope for a docs framework specifically - the drift problem really lives one layer up, in the CI/build pipeline diffing docs against the actual API surface, not in how the markdown gets served. skills as the bridge makes sense, thanks for the link, going to read through the ai config page properly
Congratulations on the launch! The service can integrate with Claude Code, right?
Aymo AI
Amazing stuff!
We generally use Fuma or Nextra, but it seems like found a better alternative. Congrats on the launch, and thanks for building this.
Dial
Markdown-first and zero-config is the right call for docs tools - Mintlify and Fumadocs both make you fight the config at some point. Curious how Blume handles versioned docs (multiple releases living side by side) - is that on the roadmap or are you staying intentionally minimal for now?
Meridian
Markdown-first plus one-command deploys is a strong combination. How does Blume handle larger documentation sets with versioned APIs and shared content across multiple products?