Transcribe podcasts using OpenAI's free Whisper language model

README.md

Podcast transcribers

Using OpenAI's whisper language model.

Usage

Make sure you have docker installed.

./run.sh podcast_file.mp3 --model tiny --language English > transcript.txt

This will transcribe the podcast file podcast_file.mp3 using the tiny model and the English language. The transcript will be saved in transcript.txt.

You can tail (tail -f transcript.txt) the transcript file on another terminal to see the text as it is being generated:

Better accuracy

Use one of the following models based on your requirements as explained in the whisper docs:

Size	Parameters	English-only model	Multilingual model	Required VRAM	Relative speed
tiny	39 M	tiny.en	tiny	~1 GB	~10x
base	74 M	base.en	base	~1 GB	~7x
small	244 M	small.en	small	~2 GB	~4x
medium	769 M	medium.en	medium	~5 GB	~2x
large	1550 M	N/A	large	~10 GB	1x
turbo	809 M	N/A	turbo	~6 GB	~8x

Dockerfile

FROM ubuntu:latest
WORKDIR /app

# source: https://pypi.org/project/openai-whisper/
RUN apt-get update
RUN apt-get install curl git ffmpeg python3-pip python3.12-venv -y
RUN python3 -m venv .venv
RUN .venv/bin/pip install -U openai-whisper
RUN .venv/bin/pip install --upgrade --no-deps --force-reinstall git+https://github.com/openai/whisper.git
RUN .venv/bin/pip install setuptools-rust

ENV PATH="/app/.venv/bin:${PATH}"

CMD [ "/bin/bash" ]

run.sh

#!/bin/bash

set -euo pipefail

if [ "$#" -lt 1 ] || [ ! -f "$1" ]; then
    echo "Podcast Transcriber"
    echo "Usage: $0 <podcast file> [extra_args...]"
    exit 1
fi

file="$1"
shift
extra_args="$@"

docker build -t podcast-transcriber .

mkdir -p data
cp "$file" data/
filename=$(basename "$file")
docker run -it --rm -v "$(pwd)/data:/app/data" podcast-transcriber whisper "data/$filename" $extra_args
rm -rf data