Blume: A Zero-Config Docs Framework That Prioritizes AI Readiness

Blume, a new open-source documentation framework, promises to simplify the creation of AI-ready documentation sites. But does it offer enough to stand out in a crowded field of documentation tools?

At its core, Blume converts a folder of Markdown files into a fully-functional documentation site without requiring configuration. This isn't a new concept, but Blume's focus on AI readability and integration sets it apart from traditional documentation frameworks.

AI-First Documentation

Blume's standout feature is its built-in support for AI consumption of documentation. Every page can be accessed as raw Markdown by appending .md to the URL, making it easy for AI models to parse the content. The framework also generates llms.txt and llms-full.txt files, which are specifically formatted for large language models.

Perhaps most notably, Blume includes a Model Context Protocol (MCP) server out of the box. This allows AI-powered development tools like Claude Code and Cursor to directly access and query the documentation. The MCP server exposes four read-only tools:

1. search_docs
2. get_page
3. list_pages
4. get_navigation

This level of AI integration isn't typically found in documentation frameworks and could be a significant advantage for teams working with AI-assisted development tools.

Under the Hood

Blume uses Astro as its underlying engine, generating a hidden Astro project to render the documentation site. This approach allows Blume to leverage Astro's performance benefits while abstracting away the complexity:

The use of Astro means that the core theme ships without client-side JavaScript, potentially improving Core Web Vitals scores out of the box.

Comparison to Alternatives

While Blume offers some unique features, it's entering a competitive space. Here's how it stacks up against some popular alternatives:

Blume's zero-config approach and built-in AI features are its main differentiators. However, its newness means it lacks the extensive ecosystem and community support of more established tools like Docusaurus.

The Verdict: Promising, But Not for Everyone

Blume shows promise, especially for teams prioritizing AI integration in their documentation workflow. Its zero-config start and built-in AI features could save significant time and effort for the right projects.

However, Blume's youth is both an asset and a liability. While it incorporates modern features, it also lacks the maturity and extensive third-party integrations of established platforms. The requirement for Node.js 22.12 or newer may also be a barrier for some teams.

For projects that heavily leverage AI in their development process and value out-of-the-box AI readiness, Blume could be a compelling choice. For others, especially those with complex documentation needs or who require extensive customization, more established tools might still be the better option.

Ultimately, Blume represents an interesting direction for documentation tools, one that acknowledges the growing importance of AI in the development process. Whether it can grow its ecosystem and overcome the challenges of being a newcomer will determine its long-term success in this competitive space.