Librarian Try it

Memory with receipts.

Librarian finds the few conversations that matter in months of chat history. Reactor answers from them and cites where it read it. If the answer isn't there, it says so.

Try it

Built for the Reactive AI Memory Retrieval Challenge START Warsaw × Poland Tech Week 2026

Six months of chat.
One question.

In August a user has tried three Korean restaurants. By late September it’s four. In October they ask. The answer is one sentence in months of chat, next to dozens that look almost right.

  1. It's a fact, a recommendation, a count, an update or a date. Each gets its own search.

  2. Keyword and meaning search over every turn, reranked on our own machine. Reactor only sees the passages we pick.

  3. About three dated passages. Dates and day gaps are computed in code, not guessed.

  4. It reads 0.2% of the archive and cites the sessions it used.

  5. Weak evidence means no answer. A wrong answer costs more than silence.

Why memory is hard

Facts change

Three Korean restaurants in August, four in September. The latest wins, and both get cited.

Answers hide in one line

“A kitchen gadget” turns out to be an Instant Pot, mentioned once in passing.

Some answers don't exist

Two dogs appear in the history: a friend's puppy and a hospital therapy dog. Neither is theirs. Librarian says so.

When the first search misses, Reactor searches again.

Reactor gets a private link to the user's archive and runs its own follow-up searches before it gives up. Then it answers from what it found, with citations.

11 of 16 facts our first pass missed, recovered by follow-up searches in offline tests

Measured, not hoped.

Every retrieval change is scored against known answers in minutes, without calling Reactor. That's how one slow, shared model got five rounds of improvements in one night.

87.1
offline score, up from 42.9 for keyword search
146/146
test questions answered by Reactor
0.2%
of each archive read per question
13
times it said “not in memory” instead of guessing

For any assistant that has to remember.

Support

“What did we promise this customer in March?”

Health and coaching

The latest dose, not the first one.

Personal agents

Your preferences, with the receipt.

  • Cites every answer
  • Searches on local models
  • Says no when it should

See it answer.

Pick a real question from the test set and watch it search.

Open the demo