How Does Google Multisearch Combine Photos and Text?

Google Multisearch lets you search using both an image and text at the same time through Google Lens. Instead of relying on keywords alone to describe something you see, you can take or upload a photo and add a short text modifier such as "in green" or "repair part" to narrow down the results. Behind the scenes, Google's AI models process the visual features of the image and the intent of your text together to return more specific matches.

What Multisearch Is and How to Use It

Multisearch is a feature built into Google Lens on mobile devices and the Google app. To use it, you open Google Lens, either take a picture or upload one from your gallery, and then swipe up to tap the "Add to your search" button. From there, you can type a word or phrase that modifies what you are looking for.

For example, if you photograph a pair of shoes you like, adding "in green" tells Google to find similar styles in that color. If you take a picture of a broken appliance or bicycle component, adding "repair part" or "replacement" helps surface spare parts, tutorials, or product listings rather than general information about the whole item.

How the Technology Works Together

Traditional image search matches visual patterns—shapes, colors, textures, and recognizable objects—against a massive index of images. On its own, that works well for identifying an exact product or landmark, but it struggles when you want to change one attribute or ask a specific question about the object.

Multisearch bridges this gap using multimodal AI models, such as Google's Multitask Unified Model (MUM). Rather than running an image search and a text search separately and merging the lists, the system encodes both the image and the text into a shared understanding of meaning. It identifies the main subject in the photo and applies the text modifier as a constraint or direction, much like how a person would interpret "find this exact thing, but in a different color" or "show me parts for this."

Practical Ways to Use Text Modifiers

Adding a few words to a visual search is especially useful when you do not know the exact name of an item. Common modifiers include:

  • Color or material changes: Phrases like "in blue," "leather," or "wooden" help you find variations of a product you already like.
  • Parts and maintenance: Words such as "repair part," "manual," or "filter" narrow results to replacement components and how-to guides.
  • Styling and pairing: Typing "outfit ideas" or "matching rug" after photographing a single item lets you see how it can be styled with other pieces.
  • Care instructions: Uploading a photo of a plant with "care" or "sunlight" focuses the results on how to keep it healthy.

Why It Matters for Everyday Searching

Many real-world questions start with something you can see but cannot easily describe in words. By combining a photo with a brief text prompt, Multisearch removes the guesswork of trying to invent the right keywords, making it faster to find exact variations, replacement parts, and context-specific answers.