[More Model Choices, More Freedom Offline] In this version 1.2 update, we have expanded our local downloadable model library to bring you more powerful AI options: New LLM Downloads Supported: Added download options for multiple mainstream open-source large language models (such as various parameter sizes of DeepSeek, Gemma, etc.) to meet your diverse needs from casual chatting to deep text creation. Optimized Downloading & Loading: Improved the stability and checkpoint-resume functionality for local large-file downloads, making model acquisition smoother and faster. Enhanced Local Inference: Further fine-tuned the on-device inference engine to reduce heating and power consumption, rendering offline conversations more energy-efficient and responsive. Update now to unlock your exclusive, pure-local, and ultra-private personal AI assistant!

Thank you for using OfflineChatbot! In version 1.1, we have completed a major milestone refactoring, delivering a highly personalized, contextual, and true on-device AI experience. What's New in v1.1: 1. Custom System Prompts * You can now define your AI's persona, expertise, and behavioral rules in Settings. Turn your local chatbot into a coding guru, a language tutor, or a creative writing companion instantly! 2. Contextual Conversation Memory * Our offline inference engine now supports smart short-term memory! By formatting history into standard ChatML sequences, the model remembers recent conversation turns, enabling natural, coherent multi-turn chats. 3. General Settings & Language Preferences * A newly designed settings interface with persistent storage. Seamlessly switch interface languages between "English", "简体中文 (Simplified Chinese)", and "日本語 (Japanese)". Performance & Under-the-Hood Improvements: * True Local Native Inference (Android): Re-engineered the Android backend to execute high-performance GPU/CPU hybrid acceleration via the JNI C++ layer. Say goodbye to mocks and enjoy blazing-fast responses! * iOS Build Stability & Memory Safety: Refactored the Swift data layer and streamlined model load/unload operations to ensure near-zero memory footprint during long-term operation. Privacy remains our top priority—all chats and computing are executed 100% on your device, with zero data ever leaving. If you love OfflineChatbot, please support us with a 5-star rating on the App Store!