1.1.0

Searchosaurus 1.1.0

Faster

• Faster startup: index and OCR cache load in parallel, and the index is saved less often and only when it changed
• Smoother typing: snippets, previews and phrase filters do less work per keystroke, and stale async results are dropped
• The OCR worker shuts down after 60 s idle, and pdf.js documents are released properly

New

• Indexing status line, honest empty states, and feedback after "Rebuild index"
• OCR/PDF hits explain themselves with a snippet in the row and in the preview
• Selected text becomes the initial query, plus a new obsidian://searchosaurus URI
• Cmd/Ctrl+1–9 opens in the current tab, Cmd/Ctrl+Enter opens in a new tab, Cmd/Ctrl+Alt+Enter opens in a split. Shortcut hints match your platform
• Synced data.json is merged (recent opens and history from all devices are kept), and pins are saved immediately

Fixed

• word -exclude with several words or tags no longer returns the wrong notes. Exclusions are literal and apply to all fields
• Umlauts, ß and accents are matched correctly in snippets, the preview and the jump-to line
• Tag search together with a type operator works again
• Excluded folders disappear from results immediately, and running OCR/PDF jobs no longer bring them back
• Index and user data are saved on quit instead of being discarded
• OCR/PDF text that went missing is filled in again after a restart. The cache is checked first on every platform, including mobile
• PDFs too large for mobile are no longer cached as empty for all devices
• A failed OCR asset download on one device no longer switches OCR off everywhere via sync
• Changing OCR languages re-recognizes queued images with the new languages
• Edits in an open settings tab are no longer lost when settings arrive via sync
• The preview appears when the window is widened while search is open
• Fuzzy matching is stricter: it starts at 5 characters and allows at most 2 edits