Open source

Memory is designed — not dumped.

Bad memory contaminates agents. A designed memory layer is what makes them dependable — and becomes an institution's lasting memory. Ours is built in the open.

every record — linked, versioned, audited · designed memory, not a data dump

Belief Extraction Engine · MIT License

BEE — persistent, designed memory for AI agents

BEE gives agents a structured, versioned, source-linked memory that survives across sessions — the difference between an agent that compounds context and one that contaminates itself. It powers recall and the project wiki in every workflow we build.

  • Structured beliefs, versioned and source-linked
  • Scoped, expiring per-user memory
  • Built to be measured against public benchmarks
  • MIT licensed — no lock-in
Star on GitHub View repo
bee-recall · illustrative
[BEE recall — designed memory]
> [wiki, official]   SOP template v3 · approved · source-linked
> [user, scoped]     "prefers email follow-up" · expires 30d
> [pii, never-stored] name · phone · address — stripped at intake

Why open

Trust you can inspect

Inspectable

If a component is load-bearing for trust, it should be open enough to inspect.

Benchmarked

Measured against shared, public benchmarks — not internal marketing numbers.

No lock-in

Permissive licensing. Use it with us, or entirely without us.

Next

See how designed memory powers real workflows

The Ace platform Use cases