The chat can make the apps a note holds as well as write. Ask for flashcards, a quiz, fill in the blank sentences, a mind map, a diagram, a slide deck, stories, an infographic, a Brainrot video or a mini app, and the agent puts it in a note. It reads the apps already in your notes too, so it can answer questions about one and change it when you ask. It works the same in the chat on Mac, Windows, Linux, in your browser and on iPhone, iPad and Android, and like the chat itself it is part of the Plus, Pro and Lifetime plans.
Asking for one
Ask the way you would ask a person: make flashcards from this note, quiz me on chapter three, turn my launch plan into a slide deck, draw the refund steps as a flowchart. The app can come from:
- A note, the one you have open or one you name. The agent reads it first, the PDFs in it included.
- The conversation so far, like make a note and flashcards out of this chat.
- What you describe, like a quiz on the capitals of South America.
The chat writes flashcards, quizzes, fill in the blank sentences, mind maps, diagrams, slide decks and stories itself, in the language of the material, with about 10 cards, questions or sentences when you name no number. A diagram can be any kind Mermaid draws: a flowchart, a sequence diagram, a timeline, a gantt chart, a pie chart, a quadrant chart, a mind map, an ER diagram, a state diagram and more. While it works, the chat shows Creating the note app, which turns into Created the note app with the app's name under it.
Infographics, Brainrot videos and mini apps
These three are made by Grape once the app is in the note, the way they are when you add one by hand: an infographic, a Brainrot video or a mini app. Grape makes them from text the chat writes, like what the conversation covered, or from the note's own text. It takes a minute or two, and Grape tells you when it is ready, see Get notified when the AI is done. For a mini app, say what it should do: a tracker for the habits in this note, with a box for each day of the week.
Pictures
Ask for pictures, and a slide deck or a story gets them drawn for its slides by the image model switched on in Settings under Models, see Which model generates it. With no image model on, the deck or the story is made without pictures, and the chat says why.
Where it goes
When what you ask is about the note you have open, the app goes in that note. Otherwise the agent makes a new note for it, with a short heading above the app. Name a note if you want it somewhere else. Make a note and a quiz from this conversation makes one note holding both: what the conversation covered, then the quiz.
Asking about one
When the agent reads a note, it reads the apps in it: the cards, the questions, the sentences, the ideas of a mind map, the code of a diagram and the slides of a deck. Ask about them the way you ask about the rest of the note: which cards are about the worktop, what is the right answer to question 4, what does slide 3 say. The answer comes from the app itself.
Changing one
Ask for the change in your own words: add 5 harder questions to the quiz, fix the second card, make the deck dark, add a branch about the budget to the mind map, redo the infographic with this text, change the mini app to also track weekends. The agent can add to what an app holds or replace it. The chat shows Updating the note app, then Updated the note app with the app's name under it.
A card or a question that keeps its front keeps its pictures and its history in Review, so fixing the answer on a card does not start it over.
Which plans
Making and changing flashcards, quizzes, fill in the blank, mind maps, diagrams, slide decks and stories from the chat works on every plan the chat works on. The others follow the plans they need when you make them by hand:
- Infographics and mini apps need Plus, Pro or Lifetime, and an infographic also needs an image model switched on.
- Brainrot videos need Pro.
- Pictures for a deck or a story need an image model switched on, and Grape Pro Image Model needs Pro.
What it leaves to you
- Drawings, recordings, PDFs and scans. The agent cannot change what is inside them.
- Locked notes. A locked note never reaches the AI, so the apps in it are never read or changed.
- Copies. An app belongs to one note. Ask for it in another note too, and the agent makes a new one there. From then on, the two are separate.
With an agent
The agents you make make and change note apps too, while their Create and edit notes switch is on. With it off, an agent still reads the apps in the notes it can read and answers questions about them, but leaves them as they are.