NovelNotifier 1.2

What's new since 1.1

Reader mode — read chapters as clean, styled text

• Chapters now open in Reader view: the chapter's text extracted and rendered natively, with sepia/dark/light page themes, serif/sans toggle, and adjustable font size (persisted). ⋮ menu switches between Reader view and the full Web view; the choice is remembered.
• Per-page fallback: when a host defeats extraction, the reader shows the website with a note instead of failing silently.

Anti-scrape decoding — the part every other extractor fails on

• Font-substitution hosts decoded: sites that render English through custom fonts while the underlying text is exotic-codepoint gibberish (verified against storyseedling.com) are translated back to real text — both in Reader view and, in Web view, rewritten in the page itself so copy/share/select yields real text.
• Runtime cipher derivation: unknown substitution fonts are cracked on the fly by rendering glyphs in the page's own fonts and matching them — rotation-proof, works for Private-Use ladders and partial (vowels-only) scrambling. Genuine Chinese/Japanese text is never touched.
• Image-chapter detection: chapters rendered as image slices are recognized and routed to the web view with an explanation.
• Comments-proof extraction: known platform content containers are preferred (validated), so pages whose comment sections outweigh the chapter still extract the story.
• Extraction filters hidden decoy/honeypot text by computed styles, strips embedded watermark tokens, and drops copyright footers and teaser spam.

Also verified live on-device: chapter extraction and book navigation (‹ Prev / Next ›) across multiple translator hosts.

The decoding toolkit is also published standalone as font-decipher.

Full changelog: v1.1...v1.2