Memory for researchers
Tell it once. It remembers.
Most researchers explain the same things to every new chat. Say once what you work on and how you want a literature review written up, and the agent keeps it for every chat after, in a list you can read and edit.
Saved from what you say in chat, with no form to fill in
Every memory is a plain sentence you can read, edit or delete in settings
Tell it to forget something and it is gone
One switch turns memory off, and nothing about you reaches the model while it is off
Facts caught on the way past
Say you run a bakery while asking about customers. The bakery is kept and the question is answered.
Read it, fix it, delete it
The whole list lives in Settings under Personalization, as sentences rather than a black box.
Forget is a real command
Say forget that in the chat and the memory is deleted, not hidden.
Off means off
Turn memory off and no memory text goes out with a chat, on any device.
The phone knows what the laptop learned
Memories travel with your settings, so with sync on there is one list on every device.
Replies already in your voice
With your base style and standing instructions, the answer arrives the way you would have written it.
Results
Three things that stop being a problem
Feature lists are easy to write and hard to care about. This is what changes for researchers.
Stop rereading papers you already annotated
Ask what your notes say about a method and get the answer with the note it came from, ready to cite.
Turn interviews into text the same day
Record the session, get a transcript in the note, and code it while it is still fresh.
See how the threads connect
Link notes to each other and use the graph view to spot the argument forming across a year of reading.
In practice
What that looks like for researchers
A few moments in a normal week where the app earns its place.
Reading a new paper
Drop the PDF into a note, write alongside it, and tag it into the theme it belongs to.
Fieldwork day
Record on the phone, add photos, and let it all sync back to the desk copy.
Writing up
Ask for every note that touches a concept, then pull the quotes into the draft.
The rest of the app
What Else Grape Does
Every part of the app is built around the same idea: keep what you write, and hand it back when you need it.
AI Agent
Chat that reads your library and can act on it
Open
Semantic search
Search that works on what you meant
Open
Voice and transcription
Meetings and lectures become searchable text
Open
Web clipper
Save any page from Chrome into a note
Open
Share to Grape
A page from the share sheet becomes a note
Open
Flashcards
Decks built from the notes you wrote
Open
Ask what you already read.
Papers, interviews, field notes and half formed arguments in one local library you can question in plain language, months after you filed them.