MemriHub

MemriHub started as a personal tool I built because none of the existing PKM tools did exactly what I wanted. Most of them optimize for capturing and organizing. MemriHub was built around a different hypothesis: retrieval is the bottleneck, not capture.

Most people aren’t short on places to put notes. They’re short on reliable ways to get them back when they’re relevant.

The Design Constraint

Every architectural decision in MemriHub was filtered through one question: “Will this help users retrieve the right thing at the right time?”

That constraint shaped:

• The tagging system (faceted, not hierarchical — you don’t know in advance how you’ll want to slice your notes)
• The search (full-text by default, ranked by recency and connection density)
• The editor (Markdown-first, no WYSIWYG friction)
• The connection model (bidirectional links shown inline, not buried)

What I Learned

The hardest part wasn’t the retrieval engine — it was designing the capture experience to be fast enough that people would actually use it. A tool you don’t reach for when you need it is no tool at all.

I also learned that the market for opinionated PKM tools is intensely personal. What works for one workflow is unusable for another. Building something truly useful means picking a user — not trying to serve everyone.

Technical Notes

Built as a static-first Astro app with a thin local SQLite backend for search indexing. Deployed as a self-hosted option. Markdown files are the source of truth — no lock-in.