Zaktualizowano Nov 10, 2025
What’s New in ToolNeuron Beta 5.0
Core Functionality
Offline Speech‑to‑Text (STT) support for voice input.
Text‑to‑Speech (TTS) support with multiple voices (~200‑300 MB each), plus easy voice switching.
New Setup screen to configure models, voices, and preferences.
Data Hub & Data Packets: Manage datasets, knowledge packs, and metadata more efficiently.
Plugins
Web Searching Plugin added.
Email Writer Plugin added.
Web Scraping Plugin added.
Modular plugin architecture: you can enable/disable features as needed.
Model Support
Load any GGUF model (e.g., LLaMA, Mistral, Gemma, Quen).
Support for cloud models via OpenRouter, with choice between censored or uncensored variants.
Runtime switching between local and cloud models.
Performance & Usability Improvements
Improved animations and smoother UI interactions.
Modular design lets you enable only the features you need.
Faster model loading and better context‑handling.
Offline‑first architecture for enhanced privacy.
Installation Instructions
Download the latest APK from the Releases page.
Install on your Android device.
If prompted, enable “Install from Unknown Sources.”
Pro Tips
Explore the Plugin Store to add or remove features.
Try switching TTS voices or different AI models on the fly.
Use the Data Hub to attach local datasets and enhance AI context.
Known Limitations
Support for TFLite / ONNX / BIN model formats is planned for a future release.
Voice input is in a basic version; full audio pipeline improvements are coming.
In‑app updates are not yet automatic; manual updates via GitHub required.
