Workflow

Import, reference, scan, review.

Visage is designed as a short operational sequence so users can go from raw folder to a usable result set without learning a complex system.

How It Works

Designed to be understandable in under a minute.

Select an album folder.

Visage imports the photo set and prepares it for scanning without moving your originals to a remote service.

Add reference face photos.

Provide one or more images of the person you want to find so the app can build a stronger face profile.

Run the scan locally.

Each photo is processed on your machine with face detection, alignment, and matching. No internet round-trip is required.

Inspect and export the shortlist.

Review matches in a focused results view, adjust strictness, and export the images that matter.

Capabilities

Built for privacy, speed, and large real-world libraries.

Private By Default

Your photo library stays on your device.

Visage is offline-first. It does not depend on uploading albums to a web service just to search for a face.

Digestible Workflow

Import, reference, scan, review.

The app is structured as a short, step-based flow so users can move from raw folder to usable result set without detours.

Built For Volume

Handles large albums without browser-tab fragility.

Desktop-native packaging, worker-based processing, and cached thumbnails keep the experience practical for substantial libraries.

Precise Matching

Face detection and embedding search tuned for retrieval.

Visage detects faces, generates embeddings, and compares them with cosine similarity so likely matches surface quickly.

What You Get

Clear expectations before you install.

Best For

People searching personal or client image archives.

Pull every appearance of a child, partner, family member, subject, or client from a large photo collection without opening thousands of files by hand.

Platform

Desktop installers with local model execution.

Visage ships as a native desktop app for macOS and Windows, with local inference providers selected automatically where supported.

Output

A focused match set instead of a raw face database.

The goal is practical retrieval: find likely matches fast, inspect them visually, and export the photos you need.

Trust

Offline-first behavior with no account dependency.

There is no signup flow, no hosted dashboard, and no requirement to keep your library online to use the product.

Questions

Answers that remove the usual friction.

Does Visage upload my photos?

No. The product is built around local processing. Your images stay on the machine running the scan.

What makes the matches work?

Visage detects faces, aligns them, generates embeddings, and ranks similarity so you can review likely matches instead of searching manually.

Who is this best suited for?

Anyone with large personal or professional photo libraries who needs a direct way to find one person across many images.

How strict can matching be?

Adjust the confidence threshold to widen recall or tighten the shortlist for more selective review and export.