What it does

Answers drawn from your own work.

Ask what customers said about pricing, what is left before a launch, or what a long note comes down to. The assistant searches your library by meaning, reads the notes it needs, and replies in a few plain lines, with each note it relied on named in the answer.

  • Searches by meaning, not just keywords
  • Reads the notes before it answers
  • Cites every note it used, one click from opening

How it works

Every step in the open.

While it works, the chat shows what it is doing: the searches it ran and the notes it read. When an answer belongs in a note, it offers a card with the exact change. Nothing is written until you accept it, and an accepted change can be undone.

  • Each search and read listed as it happens
  • Edits, new notes, sheets, and sketches offered as cards
  • Accept, discard, or undo every change

Why it matters

Your notes stay yours.

The assistant runs on your own key from OpenAI, Anthropic, or Google, so you choose the model and pay the provider directly. The key stays on your device, the password vault is kept apart from the assistant, and privacy mode pauses it whenever you like.

Everything inside

Beside every page
Open beside notes, docs, sheets, and sketches, and ask about the one in front of you.
Across the library
Questions that span many notes, answered from the notes themselves.
Cited answers
Each note an answer draws on is named, and opens in one click.
Changes you approve
Proposed edits show as a card to accept or discard, and every change can be undone.
Your choice of model
Claude, GPT, or Gemini, on your own key, switched from the composer.
Chats that keep
Tabs for parallel chats, saved history, and memory you control in settings.

Start with the free plan.

Every part of Thinkornote, in one app, on the devices you already use.