Skip to content

Releases

v1.6.3

Choose a tag to compare

@github-actions github-actions released this 12 Oct 06:16
Release v1.6.3

v1.6.2

Choose a tag to compare

@github-actions github-actions released this 12 Oct 05:57

TinyAI v1.6.2

What's New & Improvements

  • Performance & Stability: Further improvements to overall app stability and reliability.
  • Android TV Compatibility: Continued refinements to the TV experience and remote navigation.
  • Bug Fixes: General fixes and optimizations for a smoother experience.

Architecture-Specific APK Split

To significantly reduce download sizes and optimize performance, APKs are split per target hardware architecture (ABI split). Choose the build matching your device target:

  • ARM64 (64-bit phones & tablets): app-arm64-v8a-release.apk (~179 MB)
  • ARMv7 (32-bit devices, legacy Android TV & set-top boxes): app-armeabi-v7a-release.apk (~109 MB)
  • x86_64 (Emulators & Intel/AMD Chromebooks): app-x86_64-release.apk (~183 MB)
  • Universal (All architectures bundled): app-universal-release.apk (~465 MB)

Thank you for using TinyAI! Feel free to open an issue on the repository for feedback or bug reports.

v1.6.1

Choose a tag to compare

@github-actions github-actions released this 05 Oct 04:40

TinyAI v1.6.1 Release Notes

Architecture-Specific APK Split

To significantly reduce download sizes and optimize performance, APKs are now split per target hardware architecture (ABI split). Choose the build matching your device target:

  • ARM64 (64-bit phones & tablets): app-arm64-v8a-release.apk (~179 MB)
  • ARMv7 (32-bit devices, legacy Android TV & set-top boxes): app-armeabi-v7a-release.apk (~109 MB)
  • x86_64 (Emulators & Intel/AMD Chromebooks): app-x86_64-release.apk (~183 MB)
  • Universal (All architectures bundled): app-universal-release.apk (~465 MB)

What's New & Improvements

  • TV & D-Pad Optimization: Seamless D-pad remote navigation on Android TV while preserving full touch controls on mobile.
  • Faster Native Inference: Upgraded llama.cpp C++ engine for lower latency and better multi-core CPU/GPU performance.
  • Stability & Fixes: Resolved navigation overlaps, improved theme switching, and optimized overall app stability.

Thank you for using TinyAI! Feel free to open an issue on the repository for feedback or bug reports.

v1.5.8

Choose a tag to compare

@github-actions github-actions released this 04 Oct 08:50

TinyAI v1.5.8

What's New & Improvements

  • Expanded Device Support: Added 32-bit ARM (armeabi-v7a) native library support, bringing local AI inference to Android TV set-top boxes, low-spec streaming devices, and older 32-bit hardware.
  • Broadened Hardware Compatibility: Updated manifest feature flags so non-touch screen devices (Android TV, Android XR, and set-top boxes) can run the app seamlessly.
  • Native C++ Optimizations: Fixed 32-bit C++ compiler conversion issues with llama.cpp and optimized backend initialization across low-RAM hardware.

Bug Fixes & Stability

  • Resolved compilation and runtime crashes when targeting 32-bit architectures.
  • Improved background memory handling and fallback stability for devices with ≤ 2 GB RAM.

Installation & Assets

Download app-release.apk attached below and install it directly on your device.

Thank you for using TinyAI! Feel free to open an issue on the repository for feedback or bug reports.

v1.5.5

Choose a tag to compare

@github-actions github-actions released this 01 Oct 14:43

TinyAI v1.5.5

Features & Improvements

  • GGUF Model Support: Added support for importing, downloading, loading, and chatting with GGUF models locally using llama.cpp.
  • Expanded Model Formats: Model Settings now supports .gguf, .task, and .litertlm model files.
  • Hugging Face Integration: Supported GGUF and MediaPipe .task files can now appear alongside LiteRT-LM models in remote model listings.
  • Local Inference: GGUF models run fully on-device with CPU fallback support.

Bug Fixes

  • Model Loading: Improved native backend initialization and GGUF model loading reliability.
  • Release Builds: Updated CI to fetch the pinned llama.cpp source and required Android NDK during release builds.

Installation & Assets

  • Download app-release.apk attached below and install it directly on your device.

Thank you for using TinyAI! Feel free to open an issue on the repository for feedback or bug reports.

v1.4.1

Choose a tag to compare

@github-actions github-actions released this 13 Aug 04:19

TinyAI v1.4.1

Features & Improvements

  • Model Management: Added a clearer unload and delete flow, along with clarified model storage paths for better file organization.

Bug Fixes

  • Theme Persistence: Fixed an issue where the theme selection was not saved across app restarts.

Installation & Assets

  • Download app-release.apk attached below and install it directly on your device.

Thank you for using TinyAI! Feel free to open an issue on the repository for feedback or bug reports.

v1.3.0

Choose a tag to compare

@github-actions github-actions released this 08 Aug 05:09

TinyAI v1.3.0 - Initial Release

Excited to announce the official first release of TinyAI!

Key Features

  • Local On-Device AI: Run AI models locally on your Android device without relying on cloud services.
  • Fast & Lightweight Inference: Optimized execution for smooth performance.
  • Clean & Intuitive UI: Modern layout designed for easy interaction and seamless user experience.

Installation & Assets

  • Download app-release.apk attached below and install it directly on your device.

Thank you for trying out TinyAI! Feel free to open an issue on the repository for feedback or bug reports.