Advanced
Power-user settings: API keys, processing limits, database location, optimization, and data resets.
The Advanced tab holds power-user settings and data tools. Most fields save automatically as you type.
API and processing
- Semantic Scholar API Key - optional. Adds higher rate limits for citation lookups, search, and PDF discovery. A free key is available from Semantic Scholar.
- AI Processing Batch Size - how many items are processed at once (1-50, default 5). Lower values use less memory.
- Maximum Token Threshold - maximum tokens per request (1,000-128,000, default 4,000). Higher values allow larger documents but can cost more.
- Strip AI Thinking Process - hide the reasoning section from thinking-capable models (on by default).
- PDF Fetch Timeout - seconds to search for a PDF before falling back to the browser (10-120, default 30).
Google Scholar
- CAPTCHA Prompt for PDF Downloads - when Google Scholar shows a CAPTCHA during a PDF fetch, choose Ask each time, Always solve, or Never (skip Scholar).
Database (desktop)
- Database Location - where Sciwand stores its local database and cache. Change Directory lets you move it; you can Move Data or Start Fresh. The app quits after migrating, so restart it manually.
- Optimize Databases - reclaims disk space by defragmenting the database files. No data is deleted.
Danger zone
- Clear Sciwand Data - permanently delete selected data: reference-manager data (Zotero, Mendeley, EndNote, Bookends), custom columns, items and documents, or vector embeddings. Cleared reference-manager data needs a re-sync. Tick what to remove, then Delete Selected (you must confirm).
- Full Reset - delete all cached data and require a full library re-sync. The app restarts afterwards.