ytmp3-cloud's Introduction

ytmp3 cloud

The cloud-based YouTube to mp3 host/converter

Overview

Prerequisites

bash
wget
docker
make
awscli (with admin account)
samcli

Setup

Choose a unique bucket name

echo ytmp3-cloud-"$(tr -dc a-z0-9 </dev/urandom | head -c 13 ; echo '')"

Set YTMP3_STORE_BUCKET_NAME in scripts/deploy.sh

Run the deploy script

make

Clean

Clean up build files during or after development

make clean

Destroy

Run the destroy script

make destroy

REST API

The REST API URL will be Ytmp3ApiUrl, output from ./scripts/deploy.sh

GET /mp3/{videoId}

400 Invalid videoId

{
    "error": "Invalid videoId"
}

400 Download failed

{
    "error": "Failed to download #########, please try again later",
}

200 Download pending

{
    "videoId": "#########",
    "status": "PENDING",
    "updatedAt": "2023-04-23T19:27:44.850Z",
    "createdAt": "2023-04-23T19:27:44.850Z"
}

200 Download complete

{
    "videoId": "#########",
    "status": "COMPLETE",
    "url": "https://##########.s3.#########.amazonaws.com/0Sl5-Cl6LOo.mp3",
    "updatedAt": "2023-04-23T19:27:44.850Z",
    "createdAt": "2023-04-23T19:27:44.850Z"
}

Contributors

Contributions are most welcome, please feel free to submit a PR!

Credits

John Van Sickle's ffmpeg static builds: https://johnvansickle.com/ffmpeg/
yt-dlp: https://github.com/yt-dlp/yt-dlp

ytmp3-cloud's People

Contributors

Watchers

ytmp3-cloud's Issues

Automatic file expiration in S3 and DynamoDB

To save on costs, it should expire files based on usage statistics .e.g., most recently used files should stay and least used files should go.

Central config json for easy deployment + options

There's some hardcoded values shared between the scripts which can be moved out into a central config. Not to mention, the .toml file could be replaced as well. Using jq we can fetch values from the json value in the bash scripts.

Static next.js Amplify frontend

This needs an amplify frontend using Next.js, automatically deployed and built into the SAM template.

Allow full youtube video urls instead of just videoIds

I believe that yt-dlp can handle this by default, so we don't need to do anything but check for nasty query parameter injections such as &list= etc.