zget
MCP server for archiving media from 600+ sites (YouTube, TikTok, etc.) and managing a local library with search, metadata, and download tools.
README
zget
The Archival Engine. Download and preserve media from YouTube, Instagram, TikTok, Reddit, X, Twitch, C-SPAN, and 600+ other sites to your own local library.

Why zget?
The internet is ephemeral. Your library isn't.
Content disappears constantly—videos get deleted, accounts get banned, platforms shut down. zget lets you build a personal archive that you control.
- Save before it's gone. That tutorial you keep referencing, that interview, that viral clip
- Share without barriers. Send videos to people who can't access the original (geo-blocks, login walls)
- Watch offline. Download for flights, road trips, anywhere with bad connectivity
- Research and journalism. Archive footage before it gets altered or removed
- Family media server. One household library accessible from every device
Requirements
- macOS (Apple Silicon or Intel) or Linux
- Python 3.10+
- uv (fast Python package manager)
- ffmpeg (for video processing)
Installation
1. Install Dependencies
# Install uv (if you don't have it)
curl -LsSf https://astral.sh/uv/install.sh | sh
# Install ffmpeg (macOS)
brew install ffmpeg
2. Clone the Repository
git clone https://github.com/bubroz/zget.git
cd zget
3. First Run (Setup)
# This installs all Python dependencies automatically
./zget-start.command
Your browser will open to http://localhost:9989. You're ready to archive!
Quick Start (Robust Launcher)
We recommend using the included Launcher Script for daily use.
1. Daily Use
Double-click zget-start.command in Finder.
This will:
- Open a terminal window (keep this open!)
- Start the server on port 9989
- Securely bind to your Tailscale network
- Launch your browser automatically
Why keep the window open? Running in a visible terminal ensures zget has permission to use your browser's cookies. This is critical for downloading from sites like TikTok and YouTube which block background bots.
2. Mobile Access (Tailscale)
To access your library from your phone (e.g. while away from home), use Tailscale.
-
Install Tailscale on both your Mac and your Phone.
-
Log in to the same account.
-
Visit the server IP from your phone:
http://100.x.y.z:9989(Find your Mac's Tailscale IP in the Tailscale menu bar icon)
Features
Core
- Multi-Platform Downloads: YouTube, Instagram, TikTok, Reddit, Twitch, X, and 600+ sites via yt-dlp
- Full-Text Search: Find videos by title, uploader, or description (SQLite FTS5)
- Metadata Preservation: Original titles, upload dates, view counts, descriptions
- H.264 Transcoding: Conversion for iOS/Safari compatibility (when enabled in settings)
- Duplicate Detection: By URL and file hash
Media Server Integration (Plex / Jellyfin)
- Custom Output Directory: Point downloads directly at your library folder
- Flat Structure: Skip platform subdirectories for watch-folder scanning
- NFO Sidecar Generation: Kodi-style XML metadata files
- Local Thumbnails: Poster images placed alongside videos
Verified Platforms
| Platform | Status |
|---|---|
| YouTube | ✅ Verified |
| ✅ Verified | |
| X | ✅ Verified |
| TikTok | ✅ Verified |
| ✅ Verified | |
| Twitch | ✅ Verified |
| C-SPAN | ✅ Verified |
Additional sites may work via yt-dlp but are not officially tested.
CLI Reference
Download a Video
uv run zget <url> # Download to default location
uv run zget <url> --output /path # Download to specific directory
uv run zget <url> --flat # Skip platform subdirectory
Library Commands
uv run zget search <query> # Full-text search
uv run zget stats # Library statistics
uv run zget doctor # Health check (find orphans, verify files)
uv run zget doctor --fix # Auto-fix issues
uv run zget formats <url> # List available formats without downloading
Configuration
Persistent settings stored in ~/.config/zget/config.json:
uv run zget config show # View current settings
uv run zget config set <key> <value> # Set a value
uv run zget config unset <key> # Remove a value
Common keys:
| Key | Description | Example |
|---|---|---|
output_dir |
Custom output path | /Volumes/Media/Videos |
flat |
Skip platform subdirs | true |
template |
Filename format | %(upload_date>%Y-%m-%d)s %(title)s.%(ext)s |
Plex Setup Example
uv run zget config set output_dir "/Volumes/Media/Social Videos"
uv run zget config set flat true
Videos will now download directly to your Plex library with proper metadata (NFO) and artwork generated automatically.
AI Agent Integration (MCP)
zget exposes tools for AI agents via the Model Context Protocol.
Available Tools
| Tool | Description |
|---|---|
zget_download |
Download a video from URL |
zget_search |
Full-text search the library |
zget_get_video |
Get metadata by video ID |
zget_get_local_path |
Get filesystem path for a video |
zget_extract_info |
Extract metadata without downloading |
zget_list_formats |
List available formats |
zget_check_url |
Check if URL exists in library |
zget_get_recent |
Get recently downloaded videos |
zget_get_by_uploader |
Get videos by uploader/channel |
Configuration
Add this to your agent's MCP config (e.g., claude_desktop_config.json):
{
"mcpServers": {
"zget": {
"command": "uv",
"args": ["run", "zget-mcp"],
"cwd": "/path/to/your/zget",
"env": {
"PATH": "/opt/homebrew/bin:/usr/bin:/bin"
}
}
}
}
Replace /path/to/your/zget with the actual path where you cloned the repository.
Run Standalone
uv run zget-mcp
Troubleshooting
"Connection Refused" on TikTok
If TikTok downloads fail with 0.0.0.0 or Connection Refused, check your Pi-hole or router. TikTok uses CNAME chains that may be blocked.
Required whitelist domains:
vm.tiktok.com.edgesuite.netwww.tiktok.com.edgesuite.neta2047.r.akamai.net
Server Not Starting
Check if another process is using port 9989:
lsof -i :9989
Kill the conflicting process or use a different port:
uv run zget-server --port 8080
Mobile Can't Connect
- Verify Tailscale is running:
tailscale status - Ensure both devices are logged into the same Tailscale account
- Try accessing via IP instead of hostname:
http://100.x.y.z:9989
Architecture
src/zget/
├── server/ # FastAPI backend + Web Components frontend
├── mcp/ # Model Context Protocol server
├── library/ # Video ingest pipeline (ingest, export, thumbnails)
├── queue/ # Async download queue manager
├── db/ # SQLite FTS5 database (async_store, store, models)
├── metadata/ # NFO sidecar generation
├── commands/ # CLI subcommands
├── core.py # yt-dlp wrapper (download, extract_info)
├── config.py # Centralized configuration and path constants
├── types.py # Project-local yt-dlp type aliases (YtdlpInfo, ProgressDict)
├── cookies.py # Browser cookie extraction for yt-dlp
├── net.py # Tailscale IP detection for Secure Mesh
├── health.py # Self-diagnostics and health logging
├── smokescreen.py # Platform health verification engine (yt-dlp --simulate)
├── regions.py # Geographic site filtering and regional collections
├── safe_delete.py # Trash-based file deletion (send2trash wrapper)
├── utils.py # Shared utilities (sanitize_filename, MIME)
└── cli.py # Main CLI entry point
Acknowledgments
zget is built on yt-dlp and was developed with assistance from Gemini.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.