How much memory does Tube Archivist need?
About 2 GB for a small test install and around 4 GB for a mid-to-large installation, on a dual core CPU with four threads. The stack runs three containers, Tube Archivist, Redis and Elasticsearch.
Tube Archivist watches the YouTube channels you add, downloads each new video, indexes it and keeps it in a library you search and play from your own server. Video metadata stays in its own searchable index, so a video stays findable after the channel takes it down.
The key facts for Tube Archivist.
| Requirement | Tube Archivist |
|---|---|
| Price | Free |
| Licence | GPL-3.0 |
| Web UI | Yes |
| Minimum hardware | 2 cores, 4 threads · 2 GB RAM |
| Docker Containers | 3 (app + Redis, Elasticsearch) |
| Database | Elasticsearch |
| Architecture | x86 + ARM |
| Tech stack | Python |
| Docker | Yes |
About 2 GB for a small test install and around 4 GB for a mid-to-large installation, on a dual core CPU with four threads. The stack runs three containers, Tube Archivist, Redis and Elasticsearch.
Into the folder you mount at /youtube on the Tube Archivist container. Artwork and other persistent data sit in /cache, and Elasticsearch keeps its index in its own data volume.
It stays in your library. The video file, its metadata and its place in the index are all on your server, so the archive outlives the source on YouTube.
Cobalt pulls the video or audio out of a YouTube, TikTok or Reddit link. Host the media downloader's processing server yourself and each request runs on your own server.
Paste a video link into this self-hosted downloader and yt-dlp saves the file on your server at the quality, format and folder you pick. Whole playlists and channels can download in one go, and subscriptions queue new uploads as they appear.