Search & filters

clipboarder's search box is a thin frontend over a SQLite FTS5 index.

How the index works

Every captured item lands in the items table plus a mirrored FTS5 virtual table items_fts. Three columns are indexed:

• content — the full clipboard payload
• preview — the single-line preview
• meta — the kind-specific tag (e.g. host for URLs, language for code, format for colors)

Triggers keep the FTS index in sync on every insert/update/delete.

What you can search for

• Words — react, auth, 2024 — case-insensitive substring match
• Multi-word — auth react — both terms must appear (AND)
• Prefix — clipboarder appends * to every token automatically, so typing anth finds anthropic.com

Searches are debounced 60 ms after the last keystroke. Results are ranked by:

1. Pinned items first
2. bm25(items_fts) score (lower is more relevant)
3. last_used_at desc

Filters

The chips below the search bar restrict by kind:

Chip	Filters to
All	Everything (default)
Pinned	Items you've starred
Text	Plain text without classification
Links	Generic URLs
Repos	GitHub / GitLab / Bitbucket / Codeberg / Gist URLs
Code	Detected as code (with language guess)
Images	Bitmap captures (PNG/screenshot)
Colors	#hex, rgb(), hsl()
Music	Spotify / Apple Music / YT Music / SoundCloud / Bandcamp
Video	YouTube / Vimeo / Twitch
PDFs	.pdf files
Emails	user@host.tld
Files	Other file paths

Combining a filter with a search query narrows further — e.g. Repos + tauri shows only repo URLs whose host/title matches tauri.

Performance

On a database with 5,000 items the cold-cache search is sub-5 ms; subsequent searches are sub-ms.