Archive Concierge
An AI librarian for your blog archive. Surfaces the right posts for the right reader.
What it does
Archive Concierge is a librarian that lives at the bottom of your blog. A reader types what they're looking for — "posts about prompt caching", "the one where you broke production", "things to read on a Sunday morning" — and it returns three to five articles from your archive, with a sentence on why each one fits.
It's built on top of Postgres with pgvector and OpenAI embeddings. Your articles get indexed once. Each query is a vector lookup plus a small synthesis step. Not a chatbot, not a search box. A recommendation, like a good librarian would give.
The point is to make your archive useful again. Most blogs have hundreds of posts and a homepage that surfaces the last five. Concierge surfaces the right five for the person reading right now.
Who it's for
- Writers with 100+ posts whose best work is buried.
- Indie publications who want a smarter "related articles" without spam.
- Anyone who thinks the answer to discovery isn't infinite scroll.
Coming soon. Want to be first in line?
Get notified
Subscribe to the workshop newsletter — I'll let you know the moment Archive Concierge is ready to use, plus the build notes along the way.
Part of the workshop
- DripperyLive
Email drip series for indie writers. Plain HTML, no bloat.
- Article ForgeInternal
Internal content pipeline I use daily.
- SubhookBuilding
Inbound email orchestration for writers who outgrew Zapier.
- GrowNoteInternal
Internal hub for my Substack growth sprint — daily Notes mechanics + sprint milestones.
- FramelockBuilding
Lock a face into any AI-generated frame.