Why Sombra stores knowledge as plain markdown

Sombra is opinionated about content format: everything is stored as clean markdown. No proprietary block types, no rich widget embeds, no WYSIWYG formatting that only renders inside one application. This isn't a limitation — it's a deliberate design choice with consequences for both humans and AI.

Less formatting, more thinking

Research into knowledge management tools consistently shows that feature-rich editors create a specific failure mode: people spend more time formatting their notes than writing them. The more options an editor presents — toggle blocks, multi-column layouts, inline databases, callout styles — the more cognitive load shifts from "what do I know?" to "how should I organise this?"

The result is a kind of productive procrastination where building the system feels like using the system. Templates get designed, dashboards get polished, and the actual knowledge never gets written down. Constraint eliminates this failure mode. Markdown gives you headings, lists, emphasis, links, code blocks, tables, and images. That covers the vast majority of what anyone needs to communicate knowledge clearly. The missing 5% is almost always formatting theatre — visually impressive, but adding nothing to the information itself.

AI reads text, not formatting

There's a deeper reason too. When AI consumes knowledge — whether through MCP, an API, or any other integration — the first thing that happens is rich formatting gets stripped away. Toggle states disappear. Column layouts collapse. Callout boxes become plain paragraphs. Embedded widgets become dead links. All those hours spent tweaking visual presentation are discarded at the exact moment the content is consumed by the tools that need it most.

Sombra skips that step. Content is stored the way AI reads it, which means there's no lossy conversion between what you save and what your AI assistant can actually use. Every token counts in an AI context window — clean markdown is dense with meaning and free of formatting noise.

Portable by default

Markdown is the closest thing to a universal format for structured text. It renders in every code editor, every documentation platform, every AI tool, and every static site generator. Content stored as markdown is never locked into Sombra — it's yours, in a format that works everywhere, with no export step required.

Where rich authoring fits

Sombra isn't a word processor, and it doesn't try to be. If you need a beautifully formatted document — a proposal, a report, a specification — write it wherever you write best: Google Docs, your IDE, a dedicated writing tool. Then save the URL to Sombra. We'll extract the content, make it searchable, and make it available to your AI tools via MCP. The rich document lives where it was authored; the knowledge lives in Sombra.

This separation is intentional. Authoring tools should optimise for the writing experience. Knowledge tools should optimise for findability, portability, and AI consumption. Combining both into one application is how knowledge tools become bloated editors that are mediocre at everything.