MCP Youtube Channel Transcript Extractor

What mcp-youtube-channel-transcript-extractor MCP server does

The mcp-youtube-channel-transcript-extractor MCP server connects an MCP client to Mamba Labs’ YouTube Channel Stats Extractor actor on Apify. It accepts either a bare company domain, such as shopify.com, or a YouTube handle such as @shopify or a full channel URL.

The result is a flat, Clay-compatible row containing the resolved channel’s stable UC identifier, channel name, creation date, description, and subscriber count. A domain lookup performs discovery; a supplied handle skips that discovery step. An optional company name can improve matching and helps the identity check avoid returning an unrelated profile.

The server is read-only. It does not require a YouTube API key, but every request needs an Apify token and consumes Apify credits.

How it works

The get_youtube_channel_stats tool performs the lookup and returns platform status information alongside the extracted values. Statuses include ok, not_found, not_extractable, blocked, identity_mismatch, and skipped. Consumers should inspect the status before treating a field as available. In particular, null means the value could not be determined, while false means the check completed and found no presence.

Successful lookups are cached for seven days by default. Set skipCache to true when a fresh fetch is required. The includeFollowerCounts option defaults to true; disabling it resolves the profile URL without fetching count data and uses fewer resources.

The mcp-youtube-channel-transcript-extractor MCP server can also request transcripts for the channel’s most recent videos. Set includeTranscripts to true and choose the number of videos with transcriptVideoCount, which defaults to five. Transcript results are timestamped and returned as one row per video.

Setup and configuration

Use Node.js 18 or newer and obtain an Apify API token from the Apify account integrations page. Register the package with an MCP client using a stdio configuration similar to this:

{
  "mcpServers": {
    "mamba-youtube-channel-stats": {
      "command": "npx",
      "args": ["-y", "@mambalabsdev/mcp-youtube-channel-transcript-extractor"],
      "env": { "APIFY_TOKEN": "your-apify-token" }
    }
  }
}

The only required environment variable is APIFY_TOKEN. Apify pricing and actor usage rules apply to each call.

Tools and capabilities

The server exposes get_youtube_channel_stats. Its supported inputs are:

• company_domain: the company’s bare domain; use this or a handle.
• company_name: an optional identity and search-quality hint.
• handle: a YouTube handle or channel URL.
• includeFollowerCounts: fetch count data when enabled.
• skipCache: bypass the seven-day successful-result cache.
• includeTranscripts: request recent-video transcripts.
• transcriptVideoCount: set how many recent videos to process.

Limitations and notes

YouTube rounds subscriber counts above roughly one thousand, so the returned count may not be exact. The response flags this condition. Transcript extraction is currently constrained by YouTube: the caption list may be visible, while the content request can return an empty body without a proof-of-origin token. In that case, the server reports not_extractable; this should not be interpreted as proof that the video has no captions.

Transcript requests add fetches and therefore add usage. A blocked result may be worth retrying later, while an identity mismatch indicates that a real profile was found but did not match the requested company.