Capturing Snapshots

Snapshots are the building blocks of your Style Graph. Each one teaches Glimmer something about your taste.

Taking a snapshot

Click the capture button (camera icon) in the toolbar while viewing any page. Glimmer takes a screenshot and starts processing it in the background.

You'll see a brief confirmation animation. A badge appears on the Card View button when new snapshots are available.

What gets captured

Each snapshot stores:

Screenshot — A visual capture of the page
URL — The page address for easy revisiting
Title — The page title
Timestamp — When you captured it
AI metadata — Brand, shop, tags, and detected products (extracted in background)
Embedding — A semantic vector for search (generated automatically)

Background processing

After capture, Glimmer sends the screenshot to its AI pipeline. This happens asynchronously — you can keep browsing while it works. The AI extracts:

The brand and store name
Product details (if on a product page)
Style tags and categories
A semantic embedding for search

This metadata enriches your Style Graph and makes your snapshots searchable.

Tips for a strong Style Graph

Capture what you love, not what you're analyzing. Your graph should reflect your genuine taste. Snapshot the things that catch your eye — the pieces you'd actually wear.

Vary your captures. Don't just snapshot one brand or one category. The more diverse your captures across brands, categories, and occasions, the richer your graph becomes.

Aim for 10+ snapshots early. That's the minimum for meaningful AI insights. After that, every new capture sharpens your graph.

Capture full outfits, not just individual items. When you snapshot a styled look, Glimmer picks up co-occurrence signals between the pieces — which powers "complete the look" suggestions.

Adding comments

You can add a personal note to any snapshot after capturing it. Open the snapshot in Card View and add your comment — useful for remembering why you liked something or what occasion you had in mind.