Search that uses or connects multiple types of information, such as images and text, depending on the system.
Multimodal search uses more than one type of information, such as text and images, in finding relevant results. An image can provide visual context that is difficult to describe, while text can clarify what the person wants to know about it. Supported combinations depend on the search product.
The media do not all have to be entered at the same instant. A person might start with an image and then refine the search with words. Google Lens, for example, supports image-led searching and adding a textual description. That example does not establish that every service supports audio, video and images in the same way.
Visual similarity is useful evidence for finding candidates, but it does not verify an exact match or a diagnosis. Two objects can look alike while differing in specifications, compatibility or condition. Important conclusions still need the appropriate identifying information and source checks.
Takeaway: Provide clear, accurate media alongside useful text for your products and resources. Avoid assuming that a photo alone will reliably identify a service need or send a customer to your business.
Example
A customer uses a photograph to look for a cabinet handle and adds text about its finish. The results may help identify similar designs, but the customer still needs to confirm dimensions and mounting details before ordering a replacement.
What to check
Use images you have permission to share and consider whether they reveal private information. Check the actual product’s input options and results. Do not promise a specific ranking or customer referral from visual-search availability.
For more detail, see Search with an image on Google.