Command Palette

Search for a command to run...

UnylyUnyly
Browse all

YouTube Analytics

FreeNot checked

Read-only YouTube channel analytics and statistics via the YouTube Data API.

GitHubEmbed

About

Read-only YouTube channel analytics and statistics via the YouTube Data API.

README

YouTube MCP

About | Getting Started | Tools | Roadmap


About

YouTube MCP is a small program designed to be used by desktop AI clients such as Gemini CLI or Claude Desktop to allow AI agents to READ ONLY the stats of your YouTube Channel including analytics - such as AVD, AVP, Views, Subscribers, and more.

This program does not write to your channel,
does not touch, read, write monetization data.
All data is controlled by you, with everything working through your machine.

Getting Started

OAuth and Credentials

These are required for both the Installer or Manual Installation.

  1. Create a new application on Google Developer Console
  2. Enable the YouTube Data API
  3. Enable the YouTube Analytics API
  4. Create an OAuth Screen
    1. App Information: N/A
    2. Audience: External
    3. Contact Information: N/A
    4. Finish: Accept the agreement
  5. Create an OAuth Client
    1. Application Type: Web Application
    2. Name: Whatever you want, ie: "YouTube MCP"
    3. Authorized Redirect URIs: http://localhost:9999/callback
  6. Add Yourself as Test User
  7. Publish your app as to not have your Token-Refresh not reset after 7 days
    1. Set it as Production

Note: You should save the ClientID and the ClientSecret someplace secret, You can always create a new secret

  1. Create YouTube API Key
    1. Name: Whatever you want
    2. API Restrictions: YouTube Data API v3, YouTube Analytics API
    3. Create and save API Key
  2. Run the installer script, it will ask you to write the Client ID, Client Secret, and YouTube API Key.

Installer

This is the recommended option; You can still download without it but requires more work with the token-grabber and the YouTube MCP executables.

Running the Installer requires Admin privileges due to how windows 'sees' it, It would say no published because its built in house, If you do not trust it, just manually install it.

Manual Installation

You need GoLang for this.

For this, its recommended to download the GitHub repo and manually run the token-grabber directory,

go run .\cwd\token-getter\

Which will ask you for ClientID and ClientSecret, then it would auto-open the browser to setup the RefreshToken which is needed for the Private API,

Manually setting up the Config is simple JSON. Its located in

  • On Unix systems: $XDG_CONFIG_HOME if non-empty, else $HOME/.config.
  • On Mac OS: $HOME/Library/Application Support.
  • On Windows: %AppData%.

Create a new folder for the YouTubeMCP in HomeDirectory/YouTube-MCP/, Then create a file config.json which would have this schema:

{
  "YOUTUBE_API": "YOUTUBE_API_KEY",
  "YOUTUBE_REFRESH_TOKEN": "REFRESH FROM TOKEN-GRABBER OR INSTALLER",
  "ChannelHandle": "YOUTUBE HANDLE WITHOUT '@'",
  "YOUTUBE_CLIENT_ID": "CLIENT ID FROM GOOGLE DEVELOPER CONSOLE",
  "YOUTUBE_CLIENT_SECRET": "CLIENT SECRET FROM GOOGLE DEVELOPER CONSOLE"
}

Finally, install the version of the MCP you wish to have, an put the executable anywhere as long as the AI agent can access it, recommended same folder as config.json

Connecting to an AI

Each AI app is different but the main way to connect the MCP is just a path to the binary, for example

Mac & Linux users: Run the installer script — it will configure the correct path for your system automatically.

Claude Desktop @ Windows,
Gemini CLI @ Windows

  "mcpServers": {
    "YouTube": {
      "command": "C:\\AppData\\Roaming\\YouTube-MCP\\YouTube-MCP.exe"
    }
  },

Tools

These are the available tools that the MCP exposes for AI agents to use. The assumption is that the AI would automatically use these tools based on your request.

Examples

Use it as if AI has the context of your content with questions such as the following:

  • "Give me a detailed overview of my latest video's performance, especially the retention by the transcript."
  • "Compare my latest short to my top 3 best shorts"
  • "On my latest video, where are viewers dropping the video the most?"
  • "Can you compare my latest longform video to a popular video of the same niche?"

There will always be some prompts the AI cannot fulfil, but for what this tool is designed to do - read the data of your Channel, it should help creators analyze their videos and see where they need to improve.

API

🌐 Public: Requires the YouTube API Key Only
🔒 Private: Request the YouTube OAuth and Refresh Token

Range Format: "30" (last 30 days), "365" (last year), "lifetime" (all time), or "YYYY-MM-DD/YYYY-MM-DD" for custom ranges.
Applies to get_top_videos, get_channel_analytics, and get_video_analytics.

Tool Name Type Description Returns
auth_check 🔒 Private
🌐 Public
Allows MCP to check if this Data API and Analytics API is valid Auth Check Struct
all_videos 🌐 Public Returns a list of all public videos on the channel, including basic data. Can specify shortform, longform, or both. List of Videos: id, title, description, thumbnail, published_at, content_type
get_video 🌐 Public Gets detailed information for a single video. ID, Title, Description, Thumbnail, PublishedAt, Duration, Views, Likes, Dislikes, CommentCount
get_video_comments 🌐 Public Gets a limited number of comments from a specific video. Author, Text, LikeCount, PublishedAt, UpdatedAt, Comment ID
get_video_transcript 🌐 Public Returns a structured list of the video transcript. Array of transcript with text and timestamps
get_top_videos 🌐 Public Returns a list of top videos given a date range, limit and content_type (defaults to last 30 days, top 10, both long and short form content). Basic: Name, Description, Likes · Detailed: adds Views, Duration, PublishedAt, CommentCount
search_video 🌐 Public Searches for videos given a query (supports basic or detailed view). Array of Basic or Detailed Video Data
get_channel 🌐 Public Gets public information for a user's (self) channel. Name, Handle, Banner, Icon, Description
get_channel_analytics 🔒 Private Gets channel analytics for a date range (defaults to 30 days). Views, Watch time, AVD, CTR, Subscribers, Demographics, Traffic sources, Daily breakdown
get_video_analytics 🔒 Private Gets video-specific analytics for a date range (defaults to last 30 days). Views, Watch time, AVD, CTR, Demographics, Traffic sources, Daily breakdown, Retention Graph

Roadmap

  • Search for any YouTube channel by handle or ID

  • Get public stats for any channel (not just your own)

  • Drop API Key requirement — OAuth only

  • Caching for Performance Improvements < Not Fully Needed>

  • Roadmaps -> Allow AI to edit Context/<year>.md with tooling to read and write (must be enabled by config)

    • Allows AI to get a context of your plans for each year
  • TikTok integration (Channel stats, video analytics, top videos)


© 2026 ChezyName

from github.com/chezyname/youtube-mcp

Installing YouTube Analytics

This server has no published package — it is built from source. Open the repository and follow its README.

▸ github.com/chezyname/youtube-mcp

FAQ

Is YouTube Analytics MCP free?

Yes, YouTube Analytics MCP is free — one-click install via Unyly at no cost.

Does YouTube Analytics need an API key?

No, YouTube Analytics runs without API keys or environment variables.

Is YouTube Analytics hosted or self-hosted?

Self-hosted: the server runs locally on your machine via the install command above.

How do I install YouTube Analytics in Claude Desktop, Claude Code or Cursor?

Open YouTube Analytics on unyly.org, pick your client tab (Claude Desktop, Claude Code, Cursor) and press Install — the config is generated automatically, no JSON editing.

Related MCPs

Compare YouTube Analytics with

Not sure what to pick?

Find your stack in 60 seconds

Author?

Embed badge for your README

Browse similar

All media MCPs