Records go in
Anything a person keeps: a note, a contact, a meeting, a receipt, a link, a file. Different shapes, one store.
The graph database under Tentanote
Put records in and it catalogues them itself. It works out what relates to what from what they say — shared tags, the same people, words in common, a note that names another — groups what belongs together, and can tell you why.
Not open yet. It ships inside Tentanote first.
Running in this page: fifteen made-up records, Tentanote's weights, nothing placed by hand.
How it works
No schema to design first, no edges to draw by hand, no folders to file into. The graph is what the records already say about each other.
Anything a person keeps: a note, a contact, a meeting, a receipt, a link, a file. Different shapes, one store.
From what the records say: tags they share, the same list, the same spaces, words in common, a body that names another record.
What strong edges hold together becomes a group, named by what it shares. A record can relate to more than one; nothing is filed.
It breaks back into its reasons — “Both tagged #travel”, “Both with Dana Okafor”, “Share kyoto, march” — not a score to take on trust.
Why it matters
A graph database stores the edges you declared and hands you a query language. Vector search hands you a similarity with nothing behind it.
Tentabase is for what sits between: a mixed collection of things one person keeps, where someone has to trust a connection before acting on it. There, the explanation is the product.
Who it's for
Tentabase ships inside Tentanote first, and opens as a Swift package once it has earned it. Tentanote's waitlist hears first.
Join Tentanote's waitlist