Seeing AI: The 2025 update

The Take Picture button enables you to take a picture of the text that’s in view of the camera. That puts the text into Seeing AI’s document viewer, so you can read it at your own pace, and with greater accuracy. It also gives you access to Ask Seeing AI, so you can ask specific questions about the text.

This button is for if you want seeing AI to take a picture of the document automatically when all edges are visible. Once Seeing AI can see that the entire document is in view, it will take a picture and launch the document viewer.

Please note: If you enable Document Detection, Seeing AI won’t read out loud any text. Instead, it will guide you to ensure that all edges are visible to then take a picture of your document.

This option lets you adjust the language of text that Seeing AI detects when reading text instantly. You can select from a list of languages. When you’ve selected the language you want, a pop-up will appear asking you if you’re sure that you want to switch to that language.

This is for exactly what it says: taking a picture. When you take a picture, Seeing AI will analyse the picture and provide a description. You can then ask further questions about the picture through the Ask Seeing AI feature.

This is for switching between the front and back camera. You may, for example, want to switch to the front camera if you want to take a selfie and get a description of yourself, for example to check your outfit and/or make-up. This button can also help you know which camera you’re currently  using. If using the back camera, the button will be labelled ‘switch to front camera’. If the front camera is in use, the button will be labelled ‘switch to back camera’.

This is the button you use to have Seeing AI describe photos in your photo library. Seeing AI can also describe videos. It can describe up to ten videos per day, with the count resetting at midnight in your local time zone.

This is a feature that allows you to train Seeing AI to recognise specific individuals. You take three photos of that person, with their permission of course. Then Seeing AI will be able to identify them in photos and when using the person mode (more on that later).

This feature allows you to scan barcodes of certain products to identify them. The app will scan for barcodes automatically.

You will hear tones as the barcode becomes more in view of the camera. The more continuous the tone, the closer you are to the barcode. When Seeing AI detects a barcode, it will scan it automatically. It isn’t necessary to take the picture yourself.

This feature allows you to get descriptions of people. Bear in mind though, these descriptions are vague, and rather subjective.

Seeing AI will use the picture you take to predict the person’s age, gender and facial expression. Take into account though that it can get it very wrong, so take what it says with a pinch of salt.

There is also a ‘face recognition’ button, which allows you to train Seeing AI to recognise specific individuals. This is the exact same button that’s on the Describe Screen (see above).

This feature does what it says on the tin: helps you identify banknotes in different currencies.

Select the currency you want seeing AI to recognise, and then simply point your device at the banknote you want to identify. Seeing AI will tell you its denomination.

This feature is especially useful for currencies that do not contain tactile markings. You could use the app to identify your banknotes and then organise them into a system so that you can recognise them quickly when you come to making payments.

First off, it’s important to stress that this mode only works on certain devices: those with a LIDAR scanner.

The World channel uses your phone’s camera to scan your environment, and Seeing AI will announce if a person, or object, is present. If announcing an object, it will announce what that object is, for example ‘keyboard’ or ‘table’.

It’s best to use a headset if using this channel. The app will create a 3D map of your surroundings. For example, if the object is to the left, you will hear the announcement on the left-hand side. If using a headset that supports head tracking, Seeing AI will be able to use this to determine where the sound comes from. If the headset does not support head tracking, your phone’s camera will determine this, so it’s recommended that you hold your phone vertically with the screen towards you.

You can then view a list of objects that have been detected, and place an audio beacon on a particular object of interest so that you can locate it more easily.

This feature is for locating your own personal belongings. You will need to first train Seeing AI to recognise the items by recording four short videos that showcase the item. Seeing AI will give you audible cues when you do this.

Seeing AI also recommends you record the videos in different settings: for example with different backgrounds, and also with no other items present.

Once you’ve trained Seeing AI to find an item, you can then select that item and Seeing AI will help you find it. It will beep as it detects the object. The faster the beeps, the closer you are to the item. Bear in mind that while it can help you find your items, it can’t help you detect other objects while doing so, so it’s important to also use your own navigation skills.

This turns your device into a colour detector. Point the camera at anything and Seeing AI will announce the colours. Bear in mind though that factors like lighting conditions can impact upon successful colour detection.

Unfortunately, you can only use the back camera for colour detection. There isn’t the option to switch to the front camera.

This turns your device into a light detector. As soon as you activate this mode, Seeing AI will start detecting light. You’ll hear a continuous tone as soon as you select the Light Mode. The higher the pitch of this tone, the more light detected.

Siri Shortcuts allow you to use Siri to go straight to a feature of Seeing AI. When you activate Siri, you can just say the defined phrase and you’ll go straight to that feature.

To set up Siri Shortcuts, go to the Menu from anywhere in Seeing AI. Then select Settings, followed by Configure Siri Shortcuts.

If you set up the Shortcuts of Recognise Document and Recognise Short Text before the lay-out change, these now take you to the Read screen. All the other Siri shortcuts set up previously will still work in the same way they used to. If you are setting up Siri Shortcuts for the first time, you will not have the option to set up Recognise Document or Recognise Short Text, but there is the option of Read. The Siri command of “Open Seeing AI” will generally open the app on the Read screen.

Remember to change the spelling of Recognise, to the English UK spelling if your Siri language is English UK!

Home Screen shortcuts allow you to open specific features straight from your Home Screen. You can set up to four home screen shortcuts.

You then access these by long pressing on ‘Seeing AI’ to bring up the menu. If you’re a VoiceOver user, the equivalent of a long press is either double tap and hold, or triple tap, depending on what you prefer.

To configure Home Screen Shortcuts, navigate to the Menu from anywhere in Seeing AI. Then, select Settings, followed by Home Screen Shortcuts.