Installation & Setup Guide
Hardware acceleration, Termux setup, and dependency management
Target-Aware Automated Installation
amv-stt provisions native ARM64 Whisper binaries and system dependencies out-of-the-box. Neural speaker diarization models can be installed selectively to preserve mobile storage.
1. Automated Package Installation
Python Edition (PyPI):
# 1. Install official PyPI package
pip install --upgrade amv-stt
# 2. Standard lightweight setup (Whisper STT + FFmpeg)
amv-stt install
# 3. Optional: Full Neural Diarization Provisioning (PyAnnote + CAM++ 192d)
amv-stt install --engine diarization
# Or full setup in one command:
amv-stt install --all
Node.js / TypeScript Edition (npm):
# 1. Install official global npm package
npm install -g amv-stt
# 2. Standard setup
amv-stt install
# 3. Full installation
amv-stt install --all
2. Interactive Auto-Provisioning
If you run diarization without having installed the neural models, the CLI automatically detects the missing components in interactive terminal sessions and prompts: [?] Neural Diarization models are missing. Download now? [y/N]:.
3. Android Termux Prerequisites & System Dependencies
If you are provisioning a clean Android Termux environment from scratch, execute the following system package installations:
# Update Termux repository index and install native toolchain
pkg update -y && pkg install -y \
python \
nodejs-lts \
ffmpeg \
libbluray \
libxml2 \
cmake \
make \
clang \
git \
termux-api
4. Self-Healing Diagnostic Health Check
Verify that your system environment, audio linkers, and ARM NEON binaries are fully functional:
amv-stt doctor