Builders who need more than chat recall

Notes and documents as memory

Store documents, notes, and arbitrary items alongside conversations and search across them.

In short

The chat hierarchy is the default, not the cage. Memory items can be documents, notes, or arbitrary tagged blobs — search across all types or scope to one kind.

The problem

✕Your knowledge is split across a chat table, a notes store, and a document index.
✕A pure chat-memory tool cannot hold reference documents or standalone notes.
✕You want one retrieval surface over conversations and durable knowledge together.

With Memorg

✓Store documents and notes as typed memory items with tags, next to conversational exchanges.
✓Embed and index everything in the same USearch store.
✓Query across types, or scope to just documents or just notes when you need to.

How it works

1

Add memory items

Write documents and notes as typed, tagged memory items in the same store as your chats.
2

Let them be indexed

Each item is embedded and added to the USearch index automatically.
3

Search across or within types

Retrieve across everything, or filter to a type, with the same blended ranking.

Features it uses

Typed memory itemsTaggingUSearch vector indexBlended retrieval

See the full feature set or how it works.

FAQ

+ Can Memorg store things that are not chat?

Yes. Memory items can be documents, notes, or arbitrary tagged blobs. The four-level chat hierarchy is the default surface, but it does not limit what you can store.

+ Can I search only documents?

Yes. You can search across all types or scope a query to a single kind, using the same deterministic ranking.