Auto is a smart camera app. Connect to Auto to use photos from the Auto iOS App in ChatGPT. Search by content, location, time. Auto understands what is in your photos, your ChatGPT and Agents use them.
added and missing (ids that were not found; nothing changed for them). Get picture ids from list_images or search_images, and album ids from list_albums or create_album.created: false) and the pictures are added to it, so there is no need to check first. Returns the album, added (ids now in the album) and missing (ids that were not found, which were not added).get_image / get_image_bytes whenever the user asked to see one. Returns an error if the picture is missing or not owned by the authenticated user.display_images_grid. Pipe ids from list_images / search_images straight into this tool rather than narrating them in text.display_images_carousel) over listing picture metadata in text whenever the user asked to see pictures — pipe ids from list_images / search_images straight into this tool.get_photo / get_photo_image whenever the user asked to see a photo. Returns an error if the photo is missing or not owned by the authenticated user.display_photos_grid. Pipe ids from list_photos / search_photos straight into this tool rather than narrating them in text.display_photos_carousel) over listing photo metadata in text whenever the user asked to see photos — pipe ids from list_photos / search_photos straight into this tool.imageCount) and the id of its cover picture (coverImageId). To see its pictures, call list_images or search_images with its albumId.display_image instead — this tool is for when the model needs the metadata itself.display_image instead; only use this tool when the model itself needs to inspect the pixels.app), how many pictures they have (totalImages), whether photo syncing is on (syncEnabled), and setup steps (help). app is auto_photos for the synced camera roll from the Auto Photos app, or auto for the pictures the user took in the Auto app, which have no albums. help tells the user how to get Auto Photos, and installUrl is its link (a TestFlight beta or the App Store, named in installChannel). Call this first when the user asks what Auto can do, and when list_images or search_images return no pictures. When totalImages is 0 or syncEnabled is false, show the user the help steps.display_photo instead — this tool is for when the model needs the metadata itself.display_photo instead; only use this tool when the model itself needs to inspect the image pixels.imageCount) and the id of its cover picture (coverImageId). To see an album's pictures, call list_images or search_images with its albumId.albumId), date range, screenshots (screenshot) and/or a (lat, lng, radius) bounding circle. source in the result says where they came from: auto_photos (the synced camera roll) or auto (pictures taken in the Auto app, which have no albums or screenshots). The radius here only filters the newest page, so for a location question prefer search_images, which searches the whole library geographically. Returns metadata only. To show the pictures to the user, pass the resulting ids to display_images_grid (multiple) or display_image (one) rather than describing them in text.display_photos_grid (multiple photos) or display_photo (one photo) rather than describing them in text.removed and missing (ids that were not found).source in the result says where they came from: auto_photos (the synced camera roll) or auto (pictures taken in the Auto app, matched by their descriptions only, with no albums or screenshots). With query, matches the text semantically against both the pictures themselves and the descriptions and OCR text written about them, then fuses the two. Without query, a (lat, lng, radius) circle returns what was captured there. This is the complete geographic search; list_images only filters its newest page by radius. albumId narrows the search to one album, and screenshot keeps or drops screenshots. All filters compose. Returns metadata only. To show the pictures to the user, pass the resulting ids to display_images_grid (multiple) or display_image (one) rather than describing them in text.query, runs semantic vector search over photo descriptions. Without query, returns photos matching the structured filters. All filters compose. Returns metadata only. To show the photos to the user, pass the resulting ids to display_photos_grid (multiple photos) or display_photo (one photo) rather than describing them in text.imageId to go back to the default, the newest picture in the album. If the cover picture later leaves the album, the album goes back to the default. Returns the album, with the cover in coverImageId.Click any prompt to copy it.
App Stats
24
Tools
3
Prompts
ChatGPT
Platforms
Category
Knowledge LookupWorks with
Data refreshed daily
