Search and the Knowledge Base

Your library is only as useful as your ability to retrieve what’s in it. Glassy gives you two complementary search systems: full-text search for keywords and phrases, and the Knowledge Base for semantic, meaning-based discovery.

Full-text search

Press Ctrl +K (or Cmd +K on Mac) anywhere in Glassy to open the global search overlay. Type any word or phrase and press Enter .

Glassy searches across:

• Note titles and content — full-text match
• Bookmark titles and URLs — title and URL match
• Capture content — extracted text from captured pages
• Documents — document titles and content
• Tags — tag name match

Results are ranked by relevance, with recent items weighted higher.

Search operators

Narrow your search with operators:

Operator	Example	Effect
tag:	tag:meetings	Only items with the “meetings” tag
type:	type:note	Only notes (note, bookmark, capture)
collection:	collection:research	Items in the “research” collection
-tag:	-tag:archived	Exclude items with the “archived” tag

Operators combine freely:

tag:work type:note -tag:archived

This finds notes tagged “work” that are not tagged “archived”.

The Knowledge Base workspace

The Knowledge Base (in the sidebar under Organize, when enabled) is a dedicated view of everything you’ve saved — notes, bookmarks, captures, documents, and even Obsidian vault files if connected. It’s powered by a hybrid search system that combines:

1. BM25 full-text search — keyword matching, fast and precise
2. Vector semantic search — meaning-based matching, finds conceptually related content even without keyword overlap
3. Reciprocal Rank Fusion (RRF) — merges both result sets, ranked and deduplicated

How content gets indexed

When you save a bookmark, create a note, or sync a vault file, the corpus indexer automatically:

1. Generates a vector embedding (a mathematical representation of the content’s meaning)
2. Stores it in the content_embeddings table
3. Adds the text to an FTS5 full-text index for keyword search

This happens automatically — you don’t need to manually index anything.

Using the Knowledge Base

Open the Knowledge Base workspace to:

• Search across everything — type a query and get hybrid results
• Browse by source type — filter to notes, bookmarks, documents, or vault files
• Discover connections — find content you forgot you had

On-device vs. cloud search

When you open a note, the Related Notes panel shows semantically similar notes. If you see an “On Device” badge, the similarity was computed entirely in your browser using local embeddings — no data was sent to any server.

What’s next?

Learn about importing and exporting your data to move content in and out of Glassy — Markdown export, JSON export, and Obsidian import.