ChatGPT Adds Virtual Try-On and Product Favorites
publishers 25articles 29first reported 28 Sep, 17:00 UTCdeveloping since 28 Sep · 28 editions
OpenAI is rolling out two shopping features globally in ChatGPT: virtual try-on for clothing and accessories, and a way to save products users like for later reference.
When clothing appears in ChatGPT shopping results, users can select a “Try On” button and upload a selfie or full-body photo to generate an image showing how the item might look on them. Users can also upload a screenshot or other image of a product and ask ChatGPT to create a virtual try-on image. The feature uses OpenAI’s ChatGPT Images 2.5 model, which the company says offers more natural lighting, richer textures, better adherence to editing instructions and reduces image generation latency.
Favorites allows users to save products to a Library in the ChatGPT app for later reference. Generated try-on images will also be stored there. OpenAI says users can additionally describe a desired style or upload an image of an outfit and ask ChatGPT to find purchasable pieces.
The virtual try-on concept resembles a feature Google previously introduced in Search. Engadget noted that images uploaded through personal ChatGPT accounts may be used for model training by default unless users opt out.
HOW THIS STORY WAS MADE
Written from 8 articles, 3 with the publisher's own text; 8 independent newsrooms once syndicated copies count as one; this version written 42 h after the record first saw the story; the editor revised it.
- Publishers
- 25
- Source articles
- 29
- Given to the writer
- 8, 3 with the article's own text
- Independent newsrooms
- 8
- This version written
- 42 h after the record first saw the story
- Second model (editor)
- revised
- Publication gate
- passed
- Human review
- none
WHAT THE EDITOR DID
The second model checked the writer's draft against the same sources and changed it before publication. Struck words were in the draft; underlined words are its corrections.
WHAT IT FLAGGED editor revise: Two details should be tightened to match the sourced wording without changing the story’s substance. - "OpenAI is rolling out two shopping features globally in ChatGPT: virtual try-on for clothing and accessories, and a Favorites function for saving products.": The sources say Favorites is a way to save products to a Library; they do not support the phrasing 'a Favorites function' as the feature name in the same way as virtual try-on. - "The feature uses OpenAI’s ChatGPT Images 2.5 model, which the company says offers more natural lighting, richer textures, better adherence to editing instructions and faster image generation.": The sources say the model reduces image generation latency, not generally that it offers faster image generation.
THE DRAFT AND THE CHANGES
ChatGPT Adds Virtual Try-On and Product Favorites
OpenAI is rolling out two shopping features globally in ChatGPT: virtual try-on for clothing and accessories, and a Favoritesway functionto save products users like for savinglater products.reference.
When clothing appears in ChatGPT shopping results, users can select a “Try On” button and upload a selfie or full-body photo to generate an image showing how the item might look on them. Users can also upload a screenshot or other image of a product and ask ChatGPT to create a virtual try-on image. The feature uses OpenAI’s ChatGPT Images 2.5 model, which the company says offers more natural lighting, richer textures, better adherence to editing instructions and fasterreduces image generation.generation latency.
Favorites allows users to save products to a Library in the ChatGPT app for later reference. Generated try-on images will also be stored there. OpenAI says users can additionally describe a desired style or upload an image of an outfit and ask ChatGPT to find purchasable pieces.
The virtual try-on concept resembles a feature Google previously introduced in Search. Engadget noted that images uploaded through personal ChatGPT accounts may be used for model training by default unless users opt out.
Written by a language model from the sources above, then checked by a second model that may only cut, attribute or correct. How it works →