Changelog
A history of everything new, fixed, and improved in StoryFox.
v1.9.2
Character Consistency & Richer Suggestions
- AddedStory suggestions now feature 10 diverse ideas — animals, kids, robots, toys, and more
- ChangedStronger character consistency across illustrations — species and breed stay the same on every page
- ChangedSmoother typewriter animation for story suggestions
- FixedEPUB covers now display the book title and author in iBooks and other readers
v1.9.1
Maintenance Release
- ChangedBackend enhancements and bug fixes for improved stability
v1.9.0
Bug Fixes, Model Improvements & Premium Preview
- FixedVarious bug fixes and stability improvements across the generation pipeline
- ChangedImproved model adherence — AI models now follow story prompts and illustration instructions more reliably
- AddedEarly preview of StoryFox Premium mode — a new tier with advanced AI models is coming soon
v1.8.0
Prompt Suggestions & Smarter Cloud Generation
- AddedAI-generated story prompt suggestions with typewriter animation on the creation screen — tap a suggestion to start generating instantly
- AddedRecommended models section in HuggingFace text model picker highlights top-performing models like GPT-OSS 120B and DeepSeek V3
- AddedCopy Error button on error banners lets you copy full error details for easier troubleshooting
- ChangedCloud text generation runs independently of on-device Foundation Models — no more deserialization errors when using HuggingFace or other cloud providers
- FixedCreation mode tooltips (Quick, Guided, Author) now appear directly above the hovered button instead of overlapping other UI
v1.7.6
Smarter Illustrations
- AddedMulti-concept image prompting — each story detail (character, setting, action, mood) is sent as a separate concept to the diffusion model for sharper, more faithful illustrations
- AddedProgressive shedding fallback — if image generation fails, the least important concepts are dropped one by one while preserving the character's species until last
- AddedDog and cat breed-specific species words for more accurate character rendering in illustrations
- ChangedTest harness now shows concept chips under each generated image and supports editable concept fields for prompt debugging
v1.7.5
Story Info & Release Notes
- AddedStory Info sheet now shows which AI providers and models generated your story — see the text and image engines used at a glance
- AddedStory Info sheet shows the image prompt for the current page, so you can see exactly what the AI was asked to illustrate
- AddedProvider and model names are now saved with each storybook and persist when reopening from the library
- AddedSparkle update dialog now shows "What's New" release notes — users see what changed before updating
v1.7.4
Author Mode
- AddedAuthor Mode — write your own story text page-by-page, then StoryFox generates beautiful illustrations using the full AI image pipeline
- AddedCustom hover tooltips on creation mode chips with snappy glass-material styling
- ChangedCreation mode toggle and book setup controls promoted above content for easier mode switching
- ChangedLarger, bolder hero title pulled closer to the fox illustration for a tighter visual composition
v1.7.3
Export Page as Image
- AddedExport any storybook page as a high-resolution PNG image — cover, story pages, or the ending — perfect for sharing on social media or printing a favorite illustration
v1.7.2
3D Page Turn Animation
- AddedRealistic 3D page turn animation in the book reader — pages fold and lift like a real book instead of sliding between views
- AddedDirectional page turns — forward flips the current page off to reveal the next, backward folds the previous page back in from the spine
- AddedStory Info sheet in the reader toolbar — view the original concept, format, style, characters, and full story metadata at a glance
- ChangedEach page is now a self-contained book page card with its own paper background, so the entire surface rotates as one unit during turns
- ChangedOriginal story concept is now saved with each storybook and available when reopening from the library
v1.7.1
Two-Pass Story Generation
- ChangedAll text generators (Foundation Models, MLX, Cloud) now use two-pass generation — the AI writes the complete story first, then creates illustration prompts with full narrative context, producing more visually consistent and higher-quality image descriptions
- ChangedImage prompts now reference earlier and later scenes in the story, so illustrations tell a cohesive visual narrative instead of treating each page in isolation
v1.7.0
Multi-Character Intelligence & Issue Reporting
- AddedSemantic prompt analysis engine — the on-device Foundation Model now pre-processes every image prompt to extract characters, appearance, scene setting, action, and mood for smarter fallback variants
- AddedMulti-character detection — prompts with multiple characters (e.g. a fox and an owl) now have all species and appearances extracted individually, instead of just the protagonist
- AddedCharacter reference image — the cover illustration is generated first and passed as a visual reference to ImagePlayground for the opening page, improving character consistency
- AddedReport Issue button appears in the reader toolbar when illustrations are missing — uploads a diagnostic zip (story, images, logs) for investigation
- AddedTest harness for evaluating character consistency (Debug > Test Character Harness) — tests LLM output, prompt variant chains, semantic analysis, and image generation in one place
- AddedShare Feedback button in the landing page footer — submit suggestions, bug reports, or complaints directly from the website
- AddedApp preview section on the landing page with a parallax scroll effect showcasing the creation screen
- ChangedFallback prompt variants now preserve all character species — 'friendly fox and owl in a colorful storybook scene' instead of just 'friendly fox...'
- ChangedSmart character enrichment detects when species descriptors are already inline and skips the 'Featuring...' prefix to avoid duplication
- ChangedToolbar buttons in the reader use uniform dark glass capsules, fixing the system override that turned them into coral-filled pills
v1.6.0
Character Consistency & Smarter Harness
- AddedCharacter consistency across illustrations — the app now generates a character sheet and mechanically injects it into every image prompt so characters look the same on every page
- AddedGuided mode now asks about your hero's visual appearance first (colors, clothing, features) to anchor character consistency from the start
- AddedCharacter description validation — if the AI returns weak character details, the harness extracts names and appearance from image prompts as a fallback
- AddedJSON repair for truncated model output — small models that run out of tokens no longer cause hard failures
- ChangedAll text generators (Foundation Models, MLX, Cloud) now share a single unified prompt template instead of maintaining separate copies
- ChangedGuided mode Q&A answers are now organized into structured sections (Characters, Setting, Plot, Tone) instead of a flat dump, helping the small model produce better stories
- ChangedImage retry fallbacks now preserve character descriptions — shorter prompt variants only trim the scene, never the character identity
- ChangedLanding page visual refresh with updated illustrations and layout improvements
- FixedForward/back navigation buttons in the reader are now clickable across the entire circle, not just the chevron icon
v1.5.0
Modular Settings & Visual Polish
- AddedSettings redesigned with a sidebar-tabbed layout — General, On-Device, Cloud, and About tabs for easy navigation and future extensibility
- AddedPaper texture overlay adds warm, storybook-quality grain to the app background and landing page
- AddedSparkle star accents scattered around the hero illustration on the creation screen
- AddedHugging Face and OpenRouter logos replace generic icons in Cloud settings
- ChangedHugging Face OAuth sign-in is now front and center in Cloud settings — API token input collapsed as an alternative
- ChangedHero illustration enlarged for a more impactful creation screen
- Changed"Squeeze a Story" renamed to "Craft a Story" to match the StoryFox brand
- FixedAbout panel hero image no longer appears distorted
- RemovedRemoved Z-Image Turbo from curated image models — only free HuggingFace Inference models are listed
v1.4.1
Auto-Check for Updates on Launch
- FixedApp now silently checks for updates 3 seconds after launch — no more waiting for the 24-hour Sparkle schedule or manually clicking 'Check for Updates'
v1.4.0
Redesigned Creation Screen
- ChangedCreation screen redesigned with an open, cardless layout — hero illustration, gradient title with sparkle accents, and breathing room throughout
- AddedStoryFox hero illustration on the creation screen and About panel, replacing the standard app icon
- ChangedBook setup moved from a disclosure card to a compact chip that opens a popover — sits inline with Quick/Guided toggle
- ChangedSidebar streamlined — removed header branding so New Story is the first element
- FixedSidebar selection highlight now uses coral instead of system blue across all macOS accent color settings
- ChangedApp accent color set to coral globally — focus rings, toggles, and system chrome all match the StoryFox palette
v1.3.3
Fix Cloud Image Generation Routing
- FixedCloud image generation now works with all listed HuggingFace models — Z-Image Turbo, HunyuanImage 3.0, SD 3.5 Medium, and HiDream I1 Fast previously returned 404 errors
- AddedSmart inference provider routing automatically resolves the correct HuggingFace backend (hf-inference, fal-ai, replicate) for each model
- RemovedRemoved FLUX.1 Canny and FLUX.1 Depth from image model picker — these ControlNet models require a conditioning image and don't work with text-only prompts
v1.3.2
Expanded Cloud Model Selection
- AddedHugging Face text model picker now features curated top models: GPT-OSS 120B (default), GPT-OSS 20B, Qwen3 32B, DeepSeek V3, Llama 3.1 8B, and Mistral 7B
- AddedHugging Face image model picker expanded with Z-Image Turbo, HunyuanImage 3.0, Stable Diffusion 3.5 Medium, and HiDream I1 Fast alongside FLUX.1 models
- ChangedDefault Hugging Face text model changed from Llama 4 Maverick to GPT-OSS 120B for better story generation quality
v1.3.1
Improved Tap Targets
- FixedQuick/Guided mode buttons now respond to taps anywhere on the chip, not just the text
- FixedBook Setup section can be expanded by clicking anywhere on the card, not just the label
v1.3.0
StoryFox Rebrand
- ChangedStoryJuicer is now StoryFox — new name, new domain (storyfox.app), same great storybook experience
- ChangedUpdated stamp image with StoryFox branding on every exported book
- ChangedNew HuggingFace OAuth integration for seamless cloud AI sign-in
v1.2.5
Sidebar Favorites, OpenRouter, Collapsible Sections & Landing Polish
- AddedSidebar favorites — star any storybook to pin it in a dedicated Favorites section that floats above your library
- AddedCollapsible sidebar sections — Favorites and Your Storybooks can be expanded/collapsed, with state persisted across launches
- AddedDrag-to-reorder books within each sidebar section, with order persisted via SwiftData
- AddedOpenRouter provider — now accessible in Settings as a collapsible advanced section with curated model picks at the top
- AddedFinal cover rescue pass — if the cover image is still missing after all recovery attempts, a dedicated last-resort retry runs with the safest prompt variants
- ChangedBook Setup section on creation view is now collapsible with a summary line showing current page count, format, and style
- ChangedEPUB cover image now tagged with properties="cover-image" and EPUB 2 backward-compatible meta for better Kindle/older reader support
- ChangedSelected picker items now use a stronger coral tint and border for better visual contrast
- ChangedLanding page copy tightened across Hero, Features, HuggingFace, Requirements, and Safety sections for clarity and conciseness
v1.2.4
Show Active Model in Progress View
- ChangedGeneration progress now shows the actual model name (e.g., 'FLUX.1-schnell') instead of just the provider name, and reflects fallbacks in real time
v1.2.3
Dynamic Q&A Flow & Keyboard Navigation
- ChangedGuided Q&A is now dynamic — the AI decides how many questions to ask (1-3 per round) and stops when it has enough detail, instead of always running 3 fixed rounds
- AddedArrow key page navigation now works reliably in the book reader on both macOS and iOS (with external keyboard)
v1.2.2
Kid Mode Q&A — Truly Kid-Friendly
- FixedKid mode Q&A now generates truly kid-friendly questions and answers — uses few-shot examples to enforce kindergarten-level vocabulary across all 3 rounds
- ChangedKid mode round headers now say 'Hero & World', 'Adventure & Problem', 'Ending & Feelings' instead of adult terms
v1.2.1
Kid Mode Q&A Language Fix
- FixedKid mode Q&A now uses truly kid-friendly language — no more adult vocabulary like 'protagonist' or 'conflict' in questions
v1.2.0
Guided Story Creation, Audience Mode, About Panel
- AddedGuided creation mode — AI asks follow-up questions across 3 rounds (characters, plot, tone) with A/B/C suggestions to enrich your story concept before generation
- AddedKid / Adult audience toggle in Settings — adjusts Q&A question tone and story generation language level
- AddedCustom About panel with 'Made with love by Jake Rains' credit linking to jakerains.com
- ChangedCreation view now features a Quick/Guided mode toggle below the story concept input
v1.1.1
Edit & Regenerate, StoryFox Stamp, Text Cleanup
- AddedEdit button in reader toolbar — edit author, page text, or moral directly from the reader
- AddedImage regeneration with optional custom prompt — guide the AI to get the illustration you want
- AddedStoryFox library stamp on 'The End' page — a fox-on-book ink stamp appears in the reader, PDF, and EPUB exports
- FixedText formatting cleanup — strips markdown bold/italic artifacts from cloud and MLX model outputs
- FixedSqueeze a Story button now clickable across the entire pill area, not just the text
- FixedVersion numbering corrected — About screen and Sparkle updates now show the correct version
- ChangedAuto-update checks enabled on launch — no longer requires manual 'Check for Updates' click
v1.1.0
EPUB Export, Changelog & Analytics
- AddedEPUB 3.0 Fixed Layout export — read your storybooks in Apple Books, Kindle, and other EPUB readers
- AddedExport menu in the reader toolbar with both PDF and EPUB options (macOS and iOS)
- AddedChangelog page on the landing site with version history and color-coded change types
- AddedVercel Analytics custom event tracking for downloads, GitHub clicks, and HuggingFace signups
- ChangedFooter now shows a Changelog pill button and auto-syncs the version badge
v1.0.3
Landing Page & Link Previews
- AddedLanding page at storyfox.app with full design system
- AddedOpen Graph and Twitter Card meta tags for rich link previews
- AddedVercel Web Analytics integration
- ChangedFooter GitHub link upgraded to pill button style
- FixedOG image URLs now use custom domain
v1.0.2
Sparkle Auto-Update
- AddedAutomatic update checking via Sparkle 2 with EdDSA signature verification
- AddedAppcast feed hosted on GitHub for seamless version delivery
- Added"Check for Updates" menu item in the app menu
v1.0.1
Cloud & Local AI Providers
- AddedHugging Face cloud text and image generation via OAuth login
- AddedMLX Swift local model support for on-device open-weight LLMs
- AddedSettings panel with glass-morphism design and provider test buttons
- ChangedImage generation now routes through unified provider system
v1.0.0
Initial Release
- AddedOn-device story generation using Apple Foundation Models
- AddedOn-device illustration generation using ImagePlayground
- AddedFour book formats: Standard Square, Landscape, Small Square, Portrait
- AddedPrint-ready 300 DPI PDF export
- AddedSwiftData persistence for saving and reopening storybooks
- AddedPage-by-page reader with keyboard navigation and page overview grid