Pro feature — image understanding in chat is available on Pro.
Visitors can send images in the chat — a photo of a part, a screenshot of an error, a picture of the thing that's broken. On Pro, the assistant looks at them and answers, grounded in your knowledge base rather than in generic guesswork.
What is this good for?
- "What is this part?" — a photo of a component, answered with your part number, price, and stock. Combine it with the Products catalog: add reference photos and the AI can recognise your specific items rather than describing them generically.
- "Why am I seeing this?" — a screenshot of an error message.
- "Is this the right one?" — a customer holding a product, checking a model or a fit.
- "Look at the damage" — an intake photo that would otherwise take three rounds of description.
Do I need to do anything to enable it?
No configuration is needed on a Pro plan — visitors can attach images from the widget and the assistant reasons over them. What does pay off is giving it something to compare against: a Products catalog with several photos per item, at different angles and in different lighting.
Are images used as knowledge too?
Yes, in a different way. When you upload an image (.png, .jpg) as a
knowledge document, we read the text out of it with OCR and index that text. This
works on all plans and is how you get a scanned price sheet or a photographed
notice into the knowledge base. See Documents and Notes.
The Pro feature is the live case: reasoning over an image a visitor sends during a conversation.
What about privacy?
Files sent in a chat are subject to your retention window and are erased when a visitor's data is deleted. If you handle sensitive intake, read Privacy and compliance — redaction settings and the healthcare privacy profile apply to conversations that include images too.