This update improves story continuation and chapter editing: • Better continuity with established story details and chapter outlines. • New on-demand chapter review for narrative consistency and content quality. • One-tap fixes can revise a chapter from review feedback and check it again. • Clearer save and chapter actions, with review details collapsed by default. • Improved model response compatibility to reduce interruptions during generation.

This update introduces story importing: - Import TXT, Markdown, and DRM-free EPUB files - Imported stories cannot be exported; stories created entirely within the app can still be exported

Story generation improvements: automatically track character states, relationships, locations, and plot threads across chapters, with conflict confirmation when settings clash. Chapters advance by outline beats, maintain length, and reduce repetition; plus local MLX optimizations.

New Formatted Reading: auto dialogue/narration parsing, per-character colors, web-novel layout. Smart length control keeps multi-agent output within your target range. Improved chapter context & Story Lore for better serial consistency. UI fixes.

Chapter Context Strategy: switch between Short / Medium / Long-form in Settings → Story Generation; Long-form feeds structured summaries of every prior chapter to keep characters and foreshadowing consistent. MLX local model library updated: now covers 4-bit quantized models from 0.5B to 70B, adding Qwen3, Gemma 3/4, Phi-4, DeepSeek-R1 distilled, Llama 3.1/3.2/3.3 and more.

Introducing Story Briefing Q&A — let AI write stories that truly match your vision. • Smart Clarification: Share your idea, and the AI generates 3–5 clearly categorized questions (genre / protagonist / world / plot / style) to surface the key details that are easiest to overlook. • One-Tap Answers: Every question comes with ready-made options, and you can also type your own ideas freely — no need to organize your thoughts manually. • Precise Context: All answers are automatically merged into a creative context and handed to the multi-agent engine, dramatically reducing off-topic drift and broken characterization.

This update adds MLX local model support and story templates. Use compatible on-device models to help brainstorm stories, develop characters, and write chapters—or start faster with ready-made templates for different kinds of stories.

New Story Templates: choose a template before creating a story. Includes built-in genre templates (some free), plus create/edit/delete custom templates with preset agents for Premium users.

More AI Models: Added support for xAI (Grok), Mistral AI, Cohere, Tencent Hunyuan, 01.AI (Yi), MiniMax, Baichuan, and StepFun. Aggregator Platforms: Now supports SiliconFlow, OpenRouter, Groq, and Perplexity for more flexible model choices. Easier Configuration: "Get API Key" links added for all new providers to streamline setup.

Fixed issues related to In-App Purchases and improved the stability of the subscription process. Performance improvements and bug fixes.

Optimized Generation: Improved story generation algorithms for better coherence and fixed repetition issues in certain scenarios.

Easier Configuration: Added "Get API Key" links in AI Model settings to quickly access official provider websites (OpenAI, DeepSeek, Anthropic, etc.). Experience Improvements: Optimized the setup process for third-party AI services. General: Bug fixes and performance enhancements.