Grape
Help center
AI chat and agent2 min readUpdated October 1, 2026

Ask the AI about an image in a note

The AI looks at the images inside a note, so it can read a whiteboard photo or a screenshot and work with what is in it.

A photo or screenshot you added to a note is something the AI can look at. It reads the words, numbers and tables in it and works with them, so there is no need to attach the same image to the chat again.

Ask about it

Open the note that holds the image and ask in the chat, the way you ask about any note: what is written on this whiteboard, type up what I wrote in this photo of my notebook, turn the receipt into a table, make a checklist from the steps in the photo. When the AI reads the note, the images in it go to the chat with it, so the answer comes from the picture itself. Name another note and its images come along the same way when the AI reads that note.

You do not have to mention the image. When the open note is a photo of your handwriting, asking to turn what you wrote into a note is enough: the AI knows the writing is in the photo.

Step by stepOn desktop
1Open the note that holds the image and ask about it.
2The AI reads the note with its image and works from what the image shows.
Step by stepOn mobile
1With the note that holds the image open, tap the sparkles.
2Ask about the image.
3The AI reads the note with its image and works from what the image shows.

Which image is which

The images go in the order they sit in the note, so you can say the second image or the screenshot at the bottom and the AI knows which one you mean. The step where the AI read the note lists them under the note's name, Image 1, Image 2 and so on, so you can see what it looked at.

Follow up questions

The images stay with the conversation, so the next question works without reading the note again.

Which models see images

Grape Pro Chat Model and every OpenAI, Anthropic and Google model see images on both apps. On xAI, Groq, OpenRouter and local models it depends on the model: Grok 4, Llama 4, Claude, Gemini, GPT-4o and later, and models with vision in their name see them. With any other model the AI says it cannot see the image; pick one of these in the composer's model picker and ask again.

Limits

Up to six images go along each time the AI reads a note, and up to twenty in one chat; start a new chat to look at more. Large photos are scaled down before they go, which keeps the text in them readable. An image added on another device has to finish syncing before the AI can see it on this one, and an image added from a URL is downloaded from the web each time, so it has to still be online. A locked note never reaches the AI, and neither do the images inside it.

Did this not help? Contact us with the version you are on and what you were trying to do.

Try It While the Guide Is Open

Grape is free to install and needs no account, so you can follow along in the real app.