We combined an image-to-text that analyses and understands images, generating description, with a text-to-speech model to create an audio description, helping people with sight challenges.