MCP Server

OutlierKit for your AI assistant

Analysis tools for YouTubers, in any MCP client (e.g. Claude). Connect OutlierKit and ask your assistant to research any niche — it queries the OutlierKit API and returns our outlier data, channel insights, transcripts, comments, and keyword research.

Server URL https://outlierkit.com/mcp

A remote MCP server — add the URL once in your MCP client and the tools are available in every session.

Requirements

  • A paid OutlierKit plan (Pro or higher) — MCP access is included with paid plans.
  • You authenticate by signing in with Google — use the Google account tied to your OutlierKit login.

Connect it to your MCP client

  1. In your client's connector / MCP-server settings, add a remote server with the URL https://outlierkit.com/mcp (leave any auth fields blank).
  2. Connect → a secure sign-in window opens → Sign in with Google and approve.
  3. Once connected, enable the tools and ask away.

Popular clients: Claude (claude.ai) → Settings → Connectors → Add custom connector → paste the URL. Claude Code → claude mcp add outlierkit --transport http https://outlierkit.com/mcp. Cursor and other clients → add a remote MCP server with the same URL.

Authentication & privacy

The connector uses OAuth 2.1. When you connect, you sign in with Google and consent once. We read only your email address to match your OutlierKit account and apply your plan and credits.

The connector does not access your YouTube channel, your Google account, or any private Google data. It only calls the OutlierKit API on your behalf.

Tools

Ten read-only tools. Each call draws from your OutlierKit credit balance.

Outliers

outliers_search 1 credit

Find outlier videos on a topic — fast semantic search over OutlierKit's index of high-performing videos, ranked by outlier score. The default tool for outlier-video research, including 'trending' and 'recent' queries.

outliers_refresh 5 credits

A deeper, on-demand outlier analysis that surfaces additional results for a query. Use it as a fallback when outliers_search returns too few results — e.g. a niche or brand-new topic. Slower, and costs 5 credits.

Channels

channels_search 1 credit

Find outlier channels on a topic — semantic search over OutlierKit's channel index, returning matching channels plus their top outlier videos.

channels_similar 1 credit

Find channels similar to a given channel. Optionally rank results by audience size to surface channels in the same weight class as the seed.

channels_get 1 credit

Look up a single channel's full profile — stats, classification, keywords, and recent uploads.

channels_videos 1 credit

List a channel's recent uploads.

Videos

videos_get 1 credit

Look up a single video's full details, including its outlier metrics and channel context.

videos_transcript 1 credit

Get a video's transcript.

videos_comments 1 credit

Get a video's comments, sorted by top or newest.

Keywords

keywords_research 1 credit

Expand seed keywords into a ranked list of related search terms with monthly search volumes. Use minVolume and limit to narrow the results.

Credits

Every tool call costs credits from your OutlierKit balance — the same pool used by the OutlierKit app. Most tools cost 1 credit; outliers_refresh (live/deep search) costs 5. You can see your balance and top up in the OutlierKit app.

About the data

Results are OutlierKit's own analytics data — our proprietary index of outlier videos and channels, the outlier scores our engine computes, embeddings-based similarity, AI enriched profiles, and keyword datasets — derived from public YouTube data through our own pipeline. See our privacy policy.

Support

Questions or issues? Email support@outlierkit.com. For the underlying HTTP API, see the API reference.