Dev blog
Changelog
What's shipped, fixed, and changed — in the order it happened.
July 14, 2026
Multi-character "Omni" model, model health checks & notifications
- New "Omni" image model for panels with two or more characters — it conditions on several reference images at once, so each character keeps their own look. Pick it in Settings or Story Setup; single-character panels still use the default model.
- Model options now show an Online / Offline status, so you can't accidentally pick an image server that's down and only find out after generating.
- Speech and thought bubble tails can now auto-point at the character's face — using a face detector tuned for manga/anime art.
- Regenerating a single panel now lets you tick exactly which characters are in it, for when auto-detection misses one.
- Redesigned the notifications panel: cleaner layout, a "Clear all" that asks for confirmation, and new per-type preferences so you choose which notifications you get (completed actions, warnings, errors).
July 13, 2026
Character builder redesign & guide bot streaming
- Character creation now uses quick-pick pills (gender, style, age, body, hair, eyes, face, skin) instead of long dropdown lists.
- The guide bot streams its replies in as it "types", and shares occasional pro-tips plus a dedicated welcome message on the homepage.
- Story Setup now opens in Full Setup by default, with a clearer collapsible Creative Direction and Advanced Setup.
- Refreshed the sign-in/sign-up pages, fixed broken Privacy Policy/Terms links, and updated both to reflect BYOK and reference-image support.
July 11, 2026
Guide bot gets more personality
- Added a sparkle-burst animation and occasional idle quips to the guide bot launcher.
- Improved contextual tip positioning and visibility across studio pages.
July 7, 2026
Rebrand to mOhiOm
- Updated branding across the app.
- Improved gallery content loading and character preview.
- Cleaned up sidebar labels.
July 5, 2026
Landing page overhaul
- Smooth scrolling, an FAQ accordion, and new About, Services, and Showcase sections.
July 4, 2026
Faster, more reliable image storage
- Migrated generated images to Cloudflare R2 storage.
- Added auto-retry for panel generation, with user feedback and the ability to cancel.
- Character images now generate sequentially with clearer loading state.
- Panels now generate one at a time for steadier reliability on the image server.
- Pause now truly stops generation (and Resume picks up where it left off), instead of just changing the label.
- Fixed an issue where saving to cloud could remove panels saved in an earlier session.
July 1–2, 2026
Reference images & provider settings
- Added reference-image support to character design.
- Project ID is now reflected in the URL, so reloading keeps your place.
- Added configurable text-generation provider settings.
June 29–30, 2026
Print-ready exports & sharing
- Added print-ready PDF export with bleed and crop marks.
- Added QR code and social-media sharing packs.
- Introduced publish history tracking with live read stats.
- Fixed speech-bubble positions drifting between the editor and the exported file.
- Fixed occasional export failures caused by image caching.
June 25–26, 2026
EPUB export & analytics
- Added EPUB export (replacing jsPDF with pdf-lib for PDF generation).
- New rectangular layout templates.
- Introduced a client-side analytics dashboard with event tracking.
June 20–23, 2026
Dialogue editor & smarter layouts
- Shipped the dialogue editor for placing speech and thought bubbles on panels.
- Added AI layout suggestions and zoom/pan in the canvas editor.
June 12–17, 2026
Community Gallery
- Comics can now be published and read by others via the Community Gallery and a comic reader modal.
- Added character and comic ratings.
- Added smart auto-layout for panel composition.
March 24, 2026
Project kickoff
- First commit — FastAPI backend and Next.js frontend scaffold.