Changelog

A reader asked how to move articles from History to Saved. The feature did not exist yet — so we built it the same day. That, plus a batch of fixes users actually felt.

• NewSave for later — any article in your History can now be moved to Saved with one click, ready to listen whenever. Requested by one of our earliest users in the morning, live by the evening.
• FixedYouTube "Listen to Summary" now reads the actual AI summary and key takeaways. It was previously reading the raw captions transcript — no punctuation, robotic delivery. Big audible difference.
• NewVoice picker on the YouTube tab — choose from all six voices before generating, same as every other tab.
• ImprovedOne set of voice names everywhere. Sarah, James, Oliver, Emma, Michael, and Sophie now appear consistently across URL, Text, PDF, and Photo tabs — no more cryptic HD codenames on some tabs.
• FixedA failed photo OCR no longer costs you a monthly credit. If we can't read your image, the attempt is free — your 3 monthly uploads are only spent on successful extractions.
• ImprovedPhoto and PDF listens are now badged in History so you can spot them at a glance.
• PrivacyUploaded photos and PDFs are never stored — they are processed in memory to extract the text and then discarded. This was always true; now it is stated explicitly on the privacy page and inside your History.
• ImprovedHistory playback polish — play/stop from the same button, the player opens inside the history item instead of a banner at the top, and audio controls finally match dark mode.
• NewDelete your account — Settings now has a danger zone that permanently removes your data, cancels any subscription, and wipes stored audio. Your account, your call.
• NewPayment health banners — if your Pro payment fails or your access is about to lapse without renewing, the account page now tells you clearly instead of silently downgrading you.
• FixedAccount stats ("Articles listened", "Reading time saved", "This week") are now computed across your full history instead of just the visible page.

• NewSaved audio history for Pro — generated MP3s are kept for 90 days, replayable and downloadable from your account.
• FixedLong articles no longer time out during audio generation — chunks are synthesised in parallel.
• NewYour limits panel — the account page shows exactly where you stand against every daily and monthly limit, live.
• ImprovedHD voice fair-use cap (250k characters/month for Pro) introduced and reflected accurately on the pricing page.
• NewFirst-run sample article — try Sornic end-to-end without pasting anything.
• NewListening history for free accounts — metadata history is no longer Pro-only.

• NewSign in with Google — one click, works with any Google account, links automatically if you already signed up with the same email.
• NewText / Markdown reader — paste raw text or markdown and listen, no URL required.
• ImprovedFree limits are now account-aware — signing in raises your daily allowance over anonymous use.
• ImprovedFooter reorganised into responsive groups; trust and consistency fixes across SEO pages.

• NewFree AI summaries — every user now gets 3 AI analyses per day without signing up. Paste a URL, get a 3-line summary, difficulty rating, topics, and your first 2 key takeaways.
• NewYouTube AI Briefings — paste any YouTube URL and get a structured AI briefing: summary, key takeaways, topics, difficulty level, and estimated listen time. Works with podcasts, lectures, interviews, and tutorials.
• NewDeveloper API v1 + MCP server — generate summaries and audio programmatically. API keys are managed from your account.
• NewLive demo on homepage and pricing page — real Sornic output on a real article, no paste required.
• ImprovedPro upgrade prompt is now contextual — takeaways 3+, notable quotes, and action items are visible but blurred with a direct "Unlock with Pro" CTA. You see what you're paying for.
• ImprovedDark mode redesign — navy background instead of near-black, solid readable text colours, removed decorative blur orbs.
• ImprovedNew positioning: "Turn unread articles into AI briefings and audio." Sornic is an AI reading workflow, not just a TTS tool.
• ImprovedTranscript cleaning — speaker labels, music annotations, and bracketed captions are stripped before AI analysis.

• NewPro subscription launched — AI summaries, key takeaways, HD voices, MP3 downloads, PDF and photo upload, listening history, and article queue.
• NewSimplify & Listen — rewrites complex articles in plain language before reading aloud. Supports ELI5, High School, and Plain Language modes.
• NewChrome extension launched — in-page audio player, Alt+S shortcut, works on any article without leaving the page.
• ImprovedSwitched from credit pack model to clean Free + Pro subscription — one model, no confusion.

• NewPDF upload to audio — upload any PDF and listen to it with full TTS support.
• NewPhoto / OCR to audio — snap a photo of any printed text and Sornic reads it aloud.
• NewLight and dark mode — automatic system detection with manual toggle.
• NewListening history — recently played articles appear on the home screen for one-click replay, audio cached in your browser.
• ImprovedUpgraded Claude model for better TTS text cleanup — smarter handling of ads, navigation, and boilerplate.

• NewSornic launched — paste any article URL and get natural-sounding audio in seconds.
• New6 AI voices, speed control from 0.5x to 2x, MP3 downloads, and share links.
• NewTelegram bot (@SornicBot) — send any article URL and receive an MP3 directly in Telegram.
• NewBookmarklet — one-click conversion from any browser tab.