The problem
Useful knowledge is almost never in one place. On any topic — and inside any organization — it’s scattered across wikis, docs, web pages, PDFs, threads, and people’s heads. The information usually exists; what’s missing is a way to find it, trust it, and see how the pieces connect. The cost shows up as duplicated work, stale documentation, and onboarding that drags on for months.
Who it’s for — and why it matters
Anyone who has to turn sprawl into something usable: a new hire getting up to speed, a team drowning in Confluence/Notion/Slack, or — in my case — someone going deep on a subject for the fun of it. The topic is a hobby; the skill is not. Consolidating disparate sources into one organized, navigable, trustworthy reference is a core product and knowledge-management capability, and it moves straight from beer to a company’s internal docs without changing the method.
What I built
Deep, interlinked knowledge bases — “field guides” — that pull an entire domain into a single readable, navigable reference. Two so far, both built with the same repeatable method rather than hand-crafted one-offs:
- IPAs — 154 notes, 3,163 verified links
- Pour Over Coffee — 144 notes, 2,859 verified links
How it works
- Architect first. Map the domain into an information architecture — sections, categories, and atomic notes — before writing a word of content.
- Generate in parallel. Draft the notes with Claude, each one focused and cross-linked to the others with wikilinks.
- Verify integrity. Programmatically check every link — no dead ends, no orphan notes. Trust is the whole point of a reference.
- Publish. Ship it as a fast static site.
The value is in the architecture and the integrity, not the word count.
Stack & architecture
Claude for research and drafting; an Obsidian-style Markdown vault as the source of truth; a wikilink graph as the connective tissue; an automated link-verification pass; static-site generation deployed on Vercel.
How I measure success
- Coverage — does it actually span the domain end to end?
- Link integrity — share of links verified (3,163 and 2,859 verified, with zero dead ends the target). A reference people can’t trust is worthless.
- Time-to-answer — can someone find and understand a thing quickly?
- Reusability — the same pipeline produced two guides in unrelated domains, which is the real signal: the method generalizes.
Product takeaways
- This is the same problem every organization has — knowledge fragmented across systems — just dressed up in a domain I enjoy. The transferable part is the method: design the information architecture, automate the consolidation, and verify integrity so people can actually rely on the result.
- Verified links beat volume. A connected, trustworthy 150 notes is worth more than 500 disconnected ones.
- A repeatable pipeline turns “I wrote a wiki once” into “I can stand up a trustworthy knowledge base on demand” — which is the version of this skill an org actually wants.