https://github.com/mikeesto/subber

A small CLI tool for converting video & audio to a text transcription
https://github.com/mikeesto/subber

audio cli ffmpeg golang transcribe video whisper

Last synced: 4 days ago
JSON representation

A small CLI tool for converting video & audio to a text transcription

• Host: GitHub
• URL: https://github.com/mikeesto/subber
• Owner: mikeesto
• Created: 2024-04-21T09:37:35.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-08-15T04:25:27.000Z (4 months ago)
• Last Synced: 2024-10-31T19:24:35.586Z (about 2 months ago)
• Topics: audio, cli, ffmpeg, golang, transcribe, video, whisper
• Language: Go
• Homepage:
• Size: 7.81 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Subber

This is a Golang CLI tool for generating transcripts of video and audio files. It wraps FFmpeg and Whisper.

## Usage

```bash

go run *.go [--format] 

```

### Options

- `--format`: Optional flag to trigger formatting the transcript into paragraphs using basic topic analysis with TF-IDF

### Example

```bash

go run *.go --format video.mp4

```

This command will:

1. Convert `video.mp4` to a WAV

2. Transcribe the audio using Whisper

3. Format the transcript into paragraphs using topic analysis with TF-IDF

4. Save the formatted transcript to `formatted_transcript.txt`

[Some very rough benchmarks](https://gist.github.com/mikeesto/29c9cee43f95e0385b67ac83040a499e).

## Installation

I'm still working on the best way to distribute this. For now, you can clone the repository, ensure you have the required dependencies installed, and update the `WhisperPath` and `ModelPath` constants in the code to match your system.

To build the CLI:

```bash

go build -o subber

```

Then move the binary to somewhere on your PATH (e.g. `/usr/local/bin`).