Ctaylor86 MCP Video Download Server

What ctaylor86-mcp-video-download-server does

ctaylor86-mcp-video-download-server MCP server connects agent workflows to online video sources and S3-compatible object storage. It uses yt-dlp to support YouTube, Facebook, Instagram, TikTok, Twitter/X, Vimeo, Twitch, and more than 1,000 sites supported by that downloader.

The server can place several kinds of derived assets in a configured bucket:

• Video files under videos/
• MP3 audio under audio/
• Text transcripts under transcripts/
• Thumbnail images under thumbnails/

Returned results include public URLs. Depending on the operation, the response can also contain filenames, file sizes, metadata, or a transcript preview. Files are stored in the operator's S3-compatible account, and the project states that temporary files are cleaned up rather than retained permanently on the server.

How it works

The MCP client sends a source URL and, where applicable, a quality preference or language code. yt-dlp retrieves or processes the source, and the resulting asset is uploaded to the configured bucket. The server then reports the public location and operation-specific details.

The available operations are download_video_to_cloud, download_audio_to_cloud, extract_transcript_to_cloud, extract_thumbnail_to_cloud, and get_video_metadata. Metadata lookup is the non-download operation: it returns information such as title, uploader, duration, views, and description.

ctaylor86-mcp-video-download-server MCP server is described as a remote service optimized for Smithery deployment. The repository also documents a local stdio mode for development, which requires Node.js 20 or newer and a locally installed yt-dlp executable.

Setup and configuration

A deployment needs an S3-compatible endpoint, region, access key, secret key, and bucket name. Cloudflare R2 and AWS S3 are given as storage options. An optional public URL base can be used when generated links should use a custom domain.

For local operation, configure these environment variables:

• S3_ENDPOINT
• S3_REGION
• S3_ACCESS_KEY_ID
• S3_SECRET_ACCESS_KEY
• S3_BUCKET_NAME

The documented local flow is to install dependencies, build the stdio target with npm run build:stdio, and start it with npm run start:stdio. Smithery deployment uses a forked repository and storage credentials entered in the Smithery interface. The project also provides npm run dev for testing with the Smithery playground.

Tools and capabilities

download_video_to_cloud accepts a video URL and quality preference, then returns a cloud URL plus file and metadata details. download_audio_to_cloud extracts MP3 audio from a video and uploads it. extract_transcript_to_cloud accepts a URL and language code, storing clean text and returning a preview. extract_thumbnail_to_cloud uploads a thumbnail. get_video_metadata inspects a URL without downloading the media file.

Limitations and notes

The server depends on source compatibility provided by yt-dlp and on working S3-compatible credentials. Public accessibility depends on the bucket and URL configuration. Storage and hosting costs are separate considerations: the README identifies Smithery hosting and Cloudflare R2 storage as possible charges.

ctaylor86-mcp-video-download-server MCP server does not document a built-in local media library or permanent server-side storage. Its documented storage model is the configured bucket, with temporary processing files removed after use. Users should also ensure that downloading and storing material complies with the applicable source site's rules and content rights.