---
title: "Deploy Tubearchivist"
description: "Self-hosted YouTube archive: download, organize, and search your collection"
category: "Other"
url: https://railway.com/deploy/tubearchivist-2
---

# Deploy Tubearchivist

Self-hosted YouTube archive: download, organize, and search your collection

**[Deploy Tubearchivist on Railway](https://railway.com/template/tubearchivist-2)**

Machine-readable deploy manifest (JSON, validated by TemplateCI): https://railway.com/deploy/tubearchivist-2/manifest.json

- **Creator:** INAPP
- **Category:** Other

## Template content

### tubearchivist-redis

- **Source:** https://github.com/INAPP-Mobile/tubearchivist-redis

### tubearchivist-pot

- **Image:** brainicism/bgutil-ytdlp-pot-provider:1.3.2

### tubearchivist

- **Source:** https://github.com/INAPP-Mobile/tubearchivist
- **Public domain:** Yes

### tubearchivist-es

- **Source:** https://github.com/INAPP-Mobile/tubearchivist-es

## Documentation

# Deploy and Host TubeArchivist on Railway

[![Deploy on Railway](https://railway.app/button.svg)](https://railway.com/deploy/tubearchivist-2)

Self-hosted YouTube media archive server — save, organize, and search your
YouTube collection. TubeArchivist downloads videos with metadata, subtitles,
and comments, then indexes everything in Elasticsearch for fast full-text
search, all behind a clean responsive web UI.

## About Hosting TubeArchivist

This Railway template deploys the full TubeArchivist stack: the Django +
Celery + nginx main app, Elasticsearch 8.19 for search and metadata indexing,
Redis 7.4 for the task queue, and a POT-token provider companion that keeps
yt-dlp downloads working from datacenter IPs where YouTube's bot checks would
otherwise stall them. All services talk over Railway private networking, and
media persists on the main service's 50GB volume. First start takes a few
minutes while Elasticsearch initializes its index; after that, log in with
the generated credentials and add channels or video URLs to start archiving.

## Why Deploy TubeArchivist on Railway?

- **Own your archive** — videos land on your own persistent volume, immune to
  channel deletions, region blocks, or YouTube account loss.
- **Full-text search** — Elasticsearch indexes titles, descriptions, subtitles,
  and comments across your whole library.
- **Automated downloading** — subscribe to channels and playlists; new uploads
  are fetched automatically (or on-demand for single videos).
- **Bot-detection resistant** — ships with a built-in POT-token provider
  companion so yt-dlp downloads don't stall on YouTube's "confirm you're not
  a bot" checks on datacenter IPs.

## Common Use Cases

- Personal YouTube archiving: mirror channels and playlists you follow
- Offline media server: play archived videos from any browser
- Research and content-preservation projects requiring searchable archives
- Automated channel monitoring with scheduled rescans and downloads

## Dependencies for TubeArchivist Hosting

You can configure the deployment using the following environment variables
(all internal wiring is pre-configured by the template):

| Variable | Description | Default |
|---|---|---|
| `TA_USERNAME` | Initial admin username | `tubearchivist` |
| `TA_PASSWORD` | Initial admin password (auto-generated) | generated |
| `ELASTIC_PASSWORD` | Elasticsearch password (shared with ES service) | generated |
| `TA_HOST` | Public URL of this service | auto-set from Railway domain |
| `ES_URL` | Elasticsearch connection URL | auto-set from ES companion |
| `REDIS_CON` | Redis connection URL | auto-set from Redis companion |
| `POT_PROVIDER_URL` | POT-token provider for yt-dlp | auto-set from POT companion |
| `TA_MEDIA_DIR` | Media directory inside the volume | `/cache/youtube` |
| `TA_AUTO_UPDATE_YTDLP` | Auto-update yt-dlp (`release`/`nightly`/`disabled`) | `release` |
| `TZ` | Timezone (IANA name) | `UTC` |

### Deployment Dependencies

This template deploys **four services**:

1. **tubearchivist** — main app (Django + Celery + nginx), 50GB persistent
   volume at `/cache`
2. **tubearchivist-es** — Elasticsearch 8.19 (search + metadata index), volume
   at `/usr/share/elasticsearch/data`
3. **tubearchivist-redis** — Redis 7.4 (task queue + cache), volume at `/data`
4. **tubearchivist-pot** — POT-token provider (`bgutil-ytdlp-pot-provider`)
   that keeps yt-dlp downloads working from datacenter IPs

- Upstream project: [tubearchivist/tubearchivist](https://github.com/tubearchivist/tubearchivist)
- POT provider: [brainicism/bgutil-ytdlp-pot-provider](https://github.com/brainicism/bgutil-ytdlp-pot-provider)

Credentials (`TA_PASSWORD`, `ELASTIC_PASSWORD`) are auto-generated per deploy
and shared between the main app and Elasticsearch via template references —
no manual wiring needed.

### Implementation Details

In Railway, after deploying the template:

- Log in with the generated `TA_USERNAME`/`TA_PASSWORD`
- Import a YouTube cookie in Settings → Applications if downloads hit bot
  detection even with the POT provider active (datacenter IPs sometimes
  require both)
- For heavy archiving, monitor disk usage in the Railway dashboard and
  increase the volume size as needed

## Features

- Archive YouTube videos, channels, and playlists with metadata + subtitles
- Full-text search via Elasticsearch
- Automatic channel/playlist subscription monitoring
- Tagging, download queue, and playlist management
- Mobile-responsive web interface
- Built-in POT-token provider keeps downloads working on datacenter IPs

## License

GPLv3 — [TubeArchivist License](https://github.com/tubearchivist/tubearchivist/blob/master/LICENSE)

## Similar templates

- [Rocky Linux](https://railway.com/deploy/rocky-linux) — Hosted Rocky Linux 9 workspace with SSH and persistent storage. 🚀
- [Foundry Virtual Tabletop](https://railway.com/deploy/X5tR6G) — A Self-Hosted & Modern Roleplaying Platform
- [Letta Code Remote](https://railway.com/deploy/letta-code-remote) — Run a Letta Code agent 24/7. No inbound ports, just deploy.

Open this page in a browser: https://railway.com/deploy/tubearchivist-2