🚀 Product Updates

Introducing Search by Image: Find Files by Sight

Brault Team
Brault TeamThe product, design, and engineering team behind Brault.
May 3, 20264 min read

Search by image is a new way to find files in Brault. Instead of typing what you're looking for, you drop an image into the search bar and Brault returns the files in your workspace that look most like it, ranked by visual similarity. It works across photos, video frames, and PDF cover pages in the same query.

Why we built it

Creative work is full of moments where a visual is faster than a sentence. You remember the look of a shot but not the filename. A client sends a screenshot and asks "do we have anything like this?" You're trying to find every variation of a packshot that's already approved. Typing "blue tones, wide horizon, sunset" into a search bar will get you somewhere, but holding up the actual reference and saying "more like this" is closer to how creatives already think.

Search by image is built for that. Less typing, less guessing, more matching the picture in your head to the picture in your library.

How does Search by Image work?

You'll see a new image-upload icon inside the search bar. Click it, drop a photo, and Brault does the rest:

  1. Drop in a reference image. JPEG, PNG, WebP, HEIC, AVIF, and most common formats are accepted, up to 5 MB.
  2. A chip appears in the search bar with a thumbnail of your probe so you always see what you're searching with.
  3. Results render instantly in the same grid you're used to, with a similarity badge on each match so you can tell a near-perfect match from a loose one.
  4. Remove the chip to clear the search and go back to where you were.

Behind the scenes, Brault AI Search (AI Search) reads what's in your reference image and matches it against the visual fingerprints of every indexed file in your workspace. You don't see any of that. You see results.

Where it lands

A few real situations where search by image is the fastest tool in the box:

  • Designers pulling references from a moodboard. Drop the reference, find every asset in the workspace that fits the same style.
  • Photographers scanning years of shoots for a specific composition. "Wide, low horizon, warm tones" takes longer to type than to show.
  • Agencies answering client questions at speed. A client sends a screenshot, you drop it in, you have an answer in a second.
  • Brand teams governing a library. Find every variation of a logo, packshot, or campaign visual that's already in circulation.
  • Video editors finding b-roll. Search by image works on video frames too, so a still can surface the clip it came from.

It works across formats

Search by image isn't limited to photos finding photos. A single probe also surfaces:

  • Videos, when their frames look like your reference.
  • PDFs, when their cover page matches.

One reference, every visual file in your workspace. You don't have to switch search modes by file type.

Find similar files and spot duplicates

Search by image powers two more shortcuts you'll find in the file panel of any image, video, or PDF:

  • Find similar surfaces files that look like the one you're already viewing. Same logic as search by image, no upload needed. It uses the file you have open as the reference.
  • View duplicates surfaces near-exact matches: the same asset under different filenames, the same shot exported at different resolutions, the kind of accidental copies that pile up in any creative library over the years.

Both are one click. Both work cross-format. Both are the fastest way we've ever shipped to clean up a messy folder.

A few things worth knowing

  • It respects your scope. If you're inside a workspace or a folder when you run the search, results stay scoped there. No leaks across boundaries.
  • It respects permissions. You can only see files you already had access to. Search by image doesn't bypass anything.
  • Your probe stays private. The reference image you upload is processed in memory and discarded the moment the search returns. It's never stored on our servers, never indexed, never logged.
  • It works on what's been indexed. Files added before today's update are already searchable. New uploads become searchable once Brault finishes processing them, which is usually seconds.

What's next

We're already working on the next round of upgrades to search by image:

  • Paste from clipboard. Copy an image anywhere on your computer, paste it into the search bar.
  • Smarter cross-format matching for audio and document files.

And there's a bigger launch on the horizon: Brault AI Assistant. Search by image is the first piece of a much wider AI release. The Assistant will let you find, organize, tag, summarize, and work with your files just by asking, in plain language, taking AI in Brault far beyond search. The full release lands in the coming weeks. More on this soon.

If there's a workflow you'd like search by image to handle better, we want to hear about it.

Key takeaways

  • Search by image is now live in Brault. Drop a reference photo into the search bar to find visually similar files in your workspace, ranked by likeness.
  • It works across formats. One probe finds matching photos, video frames, and PDF covers in the same query.
  • Find similar and View duplicates bring the same logic to any file you have open, with one click and no upload.
  • Privacy by design. Your reference image is never stored. It exists only for the second it takes to return your results.
  • Permissions and scope are respected. You can't reach files you couldn't already see, and folder boundaries hold.

Want to see search by image in action? Book a demo or get started for free.