πΊ Full step-by-step explanation is in the video:
https://youtu.be/VIDEO_ID
(Watch this first if you want a guided walkthrough of the code and setup.)
Turn any audio/video file into:
- β
An English transcript (
.txt) - β
English subtitles (
.srt) - β
Subtitles in multiple languages (
.srtper language)
All offline, free, and no API keys.
Powered by:
- OpenAI Whisper for speech-to-text
- Argos Translate for offline translation
- Streamlit for the web UI
- π§ Speech β English text using Whisper
- π Multi-language subtitles via Argos Translate (e.g. French, Spanish, Hindiβ¦)
- π¬ Standard
.srtformat with proper timestamps - π 100% offline after first model downloads
- π₯οΈ Simple browser UI (Streamlit) β just upload & click
High-level pipeline:
- Upload audio/video file (mp3, wav, m4a, mp4, etc.)
- Whisper:
- Detects language
- Transcribes + translates to English
- Returns timestamped segments (
start,end,text)
- Subtitle builder:
- Builds English
.srtfrom Whisper segments
- Builds English
- Argos Translate:
- Translates each subtitle line English β target language
- Builds
.srtfor each selected language (same timestamps)
- Streamlit UI:
- Shows English transcript preview
- Exposes Download buttons for all generated files
.
ββ app.py # Streamlit UI (file upload, settings, downloads)
ββ transcriber.py # Core logic: Whisper, Argos, SRT generation
ββ requirements.txt # Python dependencies
ββ README.md