https://github.com/cmdoret/yt2logseq
Experiment to generate logseq notes from a youtube video using ML.
https://github.com/cmdoret/yt2logseq
logseq machine-learning note-taking video
Last synced: 9 months ago
JSON representation
Experiment to generate logseq notes from a youtube video using ML.
- Host: GitHub
- URL: https://github.com/cmdoret/yt2logseq
- Owner: cmdoret
- License: gpl-3.0
- Created: 2024-03-01T00:24:20.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2024-06-08T00:45:35.000Z (about 2 years ago)
- Last Synced: 2025-05-16T14:11:25.968Z (about 1 year ago)
- Topics: logseq, machine-learning, note-taking, video
- Language: Python
- Homepage:
- Size: 1.99 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# yt2logseq
Experiment to autogenerate [logseq](https://logseq.com/) notes from a youtube video using machine learning.
The goal of this tool is to generate logseq notes with rich metadata and AI summary from a youtube video.
It uses yt-dlp to download youtube videos, ffmpeg to extract the audio, openAI whisper to generate subtitles locally and transformers (BART models) to generate a summary of the subtitles.
In additionl the tools can read an existing logseq graph to gather all `topic::` already in use and perform zero-shot classification to assign the input video to up to 3 existing topics.
## Installation
Install with poetry
```sh
poetry install
```
## Usage
```sh
# Without logseq
python yt2logseq/main.py
# With logseq
LOGSEQ_DIR=path/to/logseq/graph python yt2logseq/main.py
```
## Example
