JevReader release notes
Notable changes to the JevReader pilot.
v0.1.0 — September 2026
Initial public pilot of the JevReader browser-based PDF reader.
- Local PDF.js rendering and text extraction with page-coordinate mapping.
- Progressive parsing in a Web Worker so reading can begin before a document finishes processing.
- Objective-aware Jev analysis over bounded batches of original source sentences.
- Semantic roles for core concepts, definitions, rules, constraints, practical insights, examples and related passage types.
- 5%, 10%, 20%, 40% and Full Focus modes with local ranking and paragraph-density controls.
- Original-sentence Essential Extract with jump-to-page behavior.
- Exact-text find over the locally parsed document.
- IndexedDB caching keyed by document hash, objective and analysis version.
- Pause, resume and retry behavior for incomplete or failed section analysis.
- Browser-side export of a new PDF with selected semantic highlights applied.
- Cloudflare Worker API with validation, origin checks, bounded payloads, timeout and concurrency protection.
Current limits: the pilot supports text-based PDFs only. OCR, semantic Find, accounts, cloud document storage and generated summaries are not included.