Download audio from m3u8 streams, transcribe with Whisper, and export to PDF, SRT, or TXT.

---

Features

	Feature	Description
1	Robust Downloading	Uses yt-dlp to handle complex HLS streams and authentication tokens
2	Accurate Transcription	Powered by OpenAI's Whisper models -- runs locally, no API keys needed
3	Multiple Output Formats	Export as PDF (with metadata), SRT subtitles, or plain text
4	Language Selection	Auto-detect or specify a language for better accuracy
5	Modern GUI	CustomTkinter interface with progress bar, output log, and cancel support
6	Dark / Light Theme	System, Dark, and Light appearance modes
7	Model Caching	Whisper models are cached in memory for faster repeated use
8	Auto Organization	Saves transcripts to transcripts/ with timestamps

---

Interface

Dark mode with output log, progress bar, format selector, and theme toggle.

---

Workflow

graph LR
    A["M3U8 URL"] -->|yt-dlp| B["MP3 Audio"]
    B -->|Whisper| C["Segments"]
    C --> D{"Format?"}
    D -->|PDF| E["Timestamped PDF"]
    D -->|SRT| F["SRT Subtitles"]
    D -->|TXT| G["Plain Text"]

Loading

---

Prerequisites

• Python 3.9+
• FFmpeg installed and on your system PATH

---

Installation

Clone and install:

git clone https://github.com/guirguispierre/m3u8-transcript.git
cd m3u8-transcript
pip install -r requirements.txt

Or install as an editable package:

pip install -e .

---

Usage

GUI (Recommended)

python3 main.py

Or explicitly:

python3 main.py --gui

CLI

Basic (PDF output):

python3 main.py "YOUR_M3U8_URL"

SRT subtitles:

python3 main.py "URL" -f srt

Plain text with custom name:

python3 main.py "URL" -f txt -o lecture_notes.txt

Medium model, French, keep audio:

python3 main.py "URL" -m medium -l fr --keep-audio

All Options

Flag	Description	Default
-f, --format	Output format: pdf, srt, txt	pdf
-m, --model	Whisper model: tiny, base, small, medium, large	base
-l, --language	ISO-639-1 language code (e.g. en, fr)	auto-detect
-o, --output	Custom output filename/path	auto-generated
--keep-audio	Keep the downloaded MP3 file	off
-v, --verbose	Enable DEBUG-level logging	off
--gui	Launch the GUI interface	--

---

Project Structure

m3u8Transcript/
├── main.py            # CLI entry point and argument parsing
├── gui.py             # CustomTkinter GUI application
├── workflow.py        # Shared download -> transcribe -> write pipeline
├── transcriber.py     # yt-dlp download + Whisper transcription
├── writers.py         # PDF, SRT, and TXT output writers
├── pdf_writer.py      # Backward-compatible PDF shim
├── logger.py          # Centralized logging configuration
├── test_pdf_gen.py    # Test suite (pytest)
├── pyproject.toml     # Package metadata and build config
├── requirements.txt   # Pinned dependencies
├── assets/
│   ├── header.png     # README banner
│   └── gui_dark.png   # GUI screenshot
└── transcripts/       # Default output directory (auto-created)

---

Running Tests

pytest -v

---

Credits

Made by Pierre Guirguis with love.