Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

tinystt — minimal whisper.cpp SRT subtitle generator

CPU-only CLI that transcribes any video/audio to sentence-level .srt using whisper.cpp tiny.en.

Download (ready builds)

Get the latest zip/tar from: https://github.com/appxa/tinystt/releases

File OS
tinystt-linux-x64.tar.gz Linux
tinystt-windows-x64.zip Windows
tinystt-macos-x64.tar.gz macOS

Each archive is only the binary. You still need ffmpeg and the model.

Linux / macOS

# 1) unpack binary onto PATH
tar -xzf tinystt-linux-x64.tar.gz   # or tinystt-macos-x64.tar.gz
mkdir -p ~/.local/bin
mv tinystt ~/.local/bin/
chmod +x ~/.local/bin/tinystt
echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.bashrc   # mac: ~/.zshrc
source ~/.bashrc

# 2) ffmpeg
# Debian/Ubuntu: sudo apt install ffmpeg
# Fedora:        sudo dnf install ffmpeg
# macOS:         brew install ffmpeg

# 3) model (once; put next to video, or set TINYSTT_MODEL)
curl -L -o ggml-tiny.en.bin \
  https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-tiny.en.bin

# 4) run
tinystt video.mp4    # → video.srt

Windows

  1. Install ffmpeg and put it on PATH.
  2. Unzip tinystt-windows-x64.zip → you get tinystt.exe.
  3. Put tinystt.exe somewhere on PATH, or keep it next to your video.
  4. Download the model once:
curl.exe -L -o ggml-tiny.en.bin https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-tiny.en.bin
  1. Run (cmd / PowerShell):
tinystt.exe video.mp4
# or: .\tinystt.exe video.mp4

Optional env vars (all OS):

TINYSTT_THREADS=2 tinystt video.mp4
TINYSTT_MODEL=/path/to/ggml-tiny.en.bin tinystt video.mp4

Default: 1 CPU thread. Model from cwd or TINYSTT_MODEL.

Build from source

Needs: C++17, cmake ≥ 3.22, g++/MSVC/clang, ffmpeg, git.

git clone https://github.com/appxa/tinystt.git
cd tinystt
cmake -B build -DCMAKE_BUILD_TYPE=Release
cmake --build build -j$(nproc)
install -m 755 build/tinystt ~/.local/bin/tinystt   # Windows: copy build\Release\tinystt.exe

whisper.cpp is fetched at build time via CMake FetchContent — no submodules.

Maintainers: tag v0.1.0 (or any v*) and push to publish release zips via CI.

License

MIT

About

speech to text by whisper model very lightweight

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages