People no longer search only by typing words. They point a phone camera at a plant, circle a jacket in a photo, or right-click an image and ask Google what it is. Until recently, that image-led traffic was invisible inside Search Console, lumped with ordinary typed searches. The Google Search Console multimodal filter changes that. Announced on September 24, 2026, it splits the Web search type in two, so you can finally see how much visibility comes from searches that start with an image rather than a keyboard.
What Is the Google Search Console Multimodal Filter?
The Google Search Console multimodal filter is a new search-type option in the Performance report. It isolates traffic from web searches in which an image, photo, or screenshot is part of the query. Four entry points are covered:
- Google Lens — pointing a camera at an object, or searching from a photo, to identify it or find similar results.
- Circle to Search on Android — circling, highlighting, or tapping something on screen to search for it without switching apps.
- Image uploads to Google Search — uploading a photo directly into the search box.
- Chrome’s “Search image with Google” — right-clicking an image on desktop (or long-pressing on mobile) to search with it.
This filter separates those visits from ordinary typed-query traffic.
Where to Find the Multimodal Filter in Search Console
- Open your property in Google Search Console and go to Performance > Search results.
- Click the Search type filter at the top of the report.
- Under Web, you will now see two options: Web: text-based and Web: multimodal.
- Select Web: multimodal to view only image-led traffic, or compare it against Web: text-based for the same date range.
- Use Export to download the filtered data and analyze it in a spreadsheet or dashboard.
The same filter is also available in the Generative AI features report. Rollout began globally on September 24, 2026, and metrics appear only for sites already receiving this traffic — an empty filter most likely means no multimodal traffic yet. Data runs from September 10 onward, so treat early comparisons as a baseline.
Web Multimodal vs Text-Based: What the Split Actually Means
A typed query tells you what someone wanted in words. An image-led search tells you what someone saw and wanted to understand, buy, or compare.
| Web: text-based | Web: multimodal | |
|---|---|---|
| How the search starts | A typed query in the search box | An image, photo, or screenshot used as part of the search |
| Entry points | Standard Google Search | Google Lens, Circle to Search, image upload, Chrome image search |
| What you learn | Which words drive impressions and clicks | Which pages surface when people search with images |
| Query data | Available in the Search results report | Not available for multimodal traffic |
| Best use | Keyword targeting and content gaps | Image SEO and visual content strategy |
Do not treat a multimodal impression as proof that a specific photo “ranked.” The filter shows only that your page appeared for an image-led search — not which image triggered it, and not whether it came from Lens or Circle to Search.
What the Google Search Console Multimodal Filter Cannot Show You
Honest limits matter more than hype, and this filter has several:
- No query data. Multimodal traffic comes with no query list.
- No source breakdown. Lens, Circle to Search, uploads, and Chrome image search are reported as one group.
- No triggering image. Google does not identify which image started the search.
- Impressions only in the Generative AI report. That report shows impressions for multimodal searches, but no clicks.
- Separate from Google Images. Web: multimodal is not the Google Images search type, which covers traffic from the Images tab. Do not merge the two.
- Measurement, not a ranking change. A jump in reported Web traffic after the rollout does not automatically mean your rankings improved.
How to Read Your Multimodal Data (Without Fooling Yourself)
Start with the Pages view, since query data does not exist for this traffic. Sort Web: multimodal pages by impressions, then compare the same pages under Web: text-based. Four patterns matter:
- High multimodal impressions, low clicks. The page surfaces but does not win the click — check the title, image quality, and whether the page answers the likely visual question fast.
- Strong in text search, near zero in multimodal. The page may rely on text alone, or its images may be hard for Google to index (see the checklist below).
- Product, tutorial, and how-to pages leading the list. These are natural fits for visual discovery.
- Sudden changes after September 10. The data starts then, so a late-September “spike” may simply be the measurement window opening.
Export both views, join them by page URL, and review monthly.
Image SEO Checklist: How to Earn More Image-Led Search Traffic
You cannot optimize for a query you cannot see, but you can make your images easy for Google to find and understand. These actions come from Google’s own guidance, echoed in Yoast’s October 2026 SEO update recap:
- Use a real, crawlable
<img>element. Google indexes images from an<img src>tag. Images delivered only as CSS backgrounds are not indexed as images. - Write descriptive filenames. Use
handmade-leather-wallet-brown.jpg, notIMG_4821.jpg. - Write helpful alt text. Describe the image in plain words, adding the relevant keyword only where it fits naturally. Alt text is also an accessibility essential.
- Use high-quality, original photos. Original photography of your actual product or subject beats blurry or stock-identical images.
- Place images near relevant text. The words around an image help Google understand what it shows.
- Keep image URLs stable. Do not change an image URL casually.
- Check that nothing blocks the image. Confirm your robots directives do not block Google from crawling the image file or its page.
For most small sites, the biggest wins are replacing placeholder filenames, adding real alt text, and moving key photos out of CSS backgrounds into proper <img> tags.
Who Benefits Most From the Multimodal Filter?
The filter is most valuable for e-commerce stores, how-to and DIY publishers, travel, food, and design sites, and local businesses — anywhere a photo of a product, project, or place can start a search. Mostly-text pages with decorative stock photos should expect modest numbers.
Frequently Asked Questions
What is the Google Search Console multimodal filter?
It is a Performance report filter that isolates web searches in which an image, photo, or screenshot is part of the search — covering Google Lens, Circle to Search on Android, image uploads, and Chrome’s “Search image with Google.”
When did Google launch multimodal reporting in Search Console?
Google announced it on September 24, 2026, on the Search Central blog, with a global rollout starting that day. Data is available from September 10 onward, and metrics appear only for sites already receiving this kind of traffic.
Does the multimodal filter show which queries or images triggered my traffic?
No. There is no query data, the sources are grouped rather than separated, and Google does not identify which image triggered a result. Use the Pages view and compare against text-based performance instead.
Is Web: multimodal the same as Google Images traffic?
No. Web: multimodal covers web searches that use an image as part of the search. The Google Images search type covers traffic from the Images tab. Report them separately.
How can I improve my site’s performance in image-led search?
Use crawlable <img> elements with descriptive filenames and helpful alt text, publish high-quality original photos, place images near relevant text, and keep image URLs stable. Avoid CSS background images for photos you want Google to index.
Conclusion: Start With the Data, Then Fix the Images
The Google Search Console multimodal filter gives small sites something they have never had before: proof of whether people are finding them by searching with images. Open the Performance report, select Web: multimodal, and compare it with Web: text-based for the same period. Note which pages appear, which ones earn clicks, and which strong text pages are missing from the image-led view — then work through the image SEO checklist on those pages first.
Image-led search will only grow as visual search becomes more natural than typing. Sites that measure it now, and fix the basics early, will be ready.
If you want help turning this kind of insight into a full visibility plan, explore TechCoff’s SEO services or get in touch for a site review. For a broader technical foundation, see the Technical SEO Checklist 2026 and our guide on how to check if an image is AI-generated.
Sources: Google Search Central: Announcing web multimodal Search performance reporting in Search Console; Yoast: The October 2026 SEO Update by Yoast recap.
