imembr

Ask your library: answers grounded in your own notes, with citations.

2026-08-12

"Ask" lets you pose a question in plain language and get an answer built from what you have actually saved — not from the open web, and not invented. It works by retrieval-augmented generation (RAG): imembr first finds the notes most relevant to your question, then asks a language model to answer using only those notes as source material.

How an answer is assembled

  • Your question is embedded and used to retrieve the closest notes and bookmarks from your library.
  • Those items are handed to the model as context, with instructions to answer from them.
  • The response cites the specific notes it drew on, so you can click through and verify.

Why citations matter

An answer you cannot trace is hard to trust. Because every response points back to the notes it used, Ask behaves less like a chatbot and more like a research assistant that shows its work. If the answer is thin, that usually means your library is thin on the topic — which is useful signal in itself.

Grounded in your data

The retrieval step only ever runs over your own account. Ask does not pull in other users' notes, and nothing you save is made public. The result is a way to interrogate your personal knowledge base that stays private to you.