Creator Intelligence
Provides tools for official YouTube research, including public channel and video lookup, bounded uploads ingestion, provenance, quota accounting, and owner-authorized YouTube Analytics/Reporting data access.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Creator Intelligenceresearch the YouTube channel MKBHD and summarize its recent uploads"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Creator Intelligence
Open-source creator research and production planning through one authenticated MCP server.
Project website · Setup · MCP tools · Verification · MIT license
Creator Intelligence gives Claude and compatible MCP clients access to official YouTube research, owner-authorized reports, original ideas and versioned production blueprints. It runs on your machine with TypeScript, Docker Compose, PostgreSQL, Redis/BullMQ, Fastify and Next.js. It is a standalone application: reviewed blueprints export as local JSON files with their revision and provenance.
Release status
Version 0.1.0 is a release candidate. The implementation and tests are public; the maintainer's deployment, credentials, private analytics and Google application are not. The complete 151-feature backlog is not implemented. Read the coverage ledger and known gaps before relying on a feature.
Public channel/video lookup, bounded uploads ingestion, provenance, quota accounting and 30-day expiry.
Read-only owner OAuth, Analytics and Reporting adapters, with explicit missing-data states.
Original idea and blueprint editing, versioning, exact-content MCP draft authorization and local blueprint JSON export.
Session/CSRF/PKCE controls, encrypted OAuth tokens, local export/deletion and encrypted original-draft backups.
Eighteen MCP tools; five governed analytics tools return a blocked response. No custom-score permission is bundled.
No scraping, quota bypass, fabricated analytics or mandatory paid cloud service. This MCP does not upload or render videos. Existing publishing tools continue separately.
Related MCP server: youtube-mcp
Quick start on Windows
Install Node.js 24.19.x, pnpm 11.25.0, PowerShell 7 and Docker Compose. Create your own Google project and enable the three YouTube APIs; see the setup guide.
git clone https://github.com/StevenBuglione/creator-intelligence.git
cd creator-intelligence
pnpm install --frozen-lockfile
./scripts/init-local.ps1 -OwnedChannelId 'YOUR_CHANNEL_ID' -GoogleProjectNumber 'YOUR_PROJECT_NUMBER'
./scripts/with-local-env.ps1 docker compose up --build -d
./scripts/copy-owner-password.ps1Replace both identity placeholders with your values. Open http://localhost:4310, paste the generated owner password, and clear the clipboard after login. Review and accept the privacy policy and terms; enable optional MCP sharing only if you want a connected AI client to receive requested records. Local drafting works before Google is connected; API-backed data requires your credentials. Installations use their own credentials and quotas. Never use the maintainer's project.
Connect one MCP server
Build the local stdio launcher, then register it with Claude Code:
pnpm build
$launcher = (Resolve-Path ./scripts/mcp-launch.ps1).Path
claude mcp add --scope user creator-intelligence -- pwsh -NoProfile -File $launcherThe Docker services must be running. An administrator may need to allow the launcher
on managed devices. The launcher loads the MCP token locally, not Google tokens.
Alternatively use the authenticated HTTP endpoint at http://localhost:4312/mcp;
register only one transport. See tool inputs, authorization and examples.
Verify locally
./scripts/with-local-env.ps1 pnpm verifyThis runs lint, TypeScript, unit/integration tests, production build and desktop/mobile
browser workflows. PostgreSQL is required. Tests use separate test/demo databases and
synthetic providers. They do not make Google API calls or prove your account connection.
Use the separate opt-in pnpm test:integration:google after owner consent.
Google approval and privacy
Open-source licensing grants rights to the code, not to Google credentials, YouTube content or additional metrics. Each operator must meet the applicable API requirements. Derived scores and extended public-statistical retention remain disabled; no approval is claimed. A personal deployment's audit is distinct from distribution of the software. Read privacy, terms, and third-party notices. Optional Claude/provider processing depends on that account's settings and is not necessarily local.
Copyright © 2026 Steven Patrick Buglione. Project code and documentation: MIT. Dependencies, service APIs and container images retain their own licenses and terms.
This server cannot be deployed
Maintenance
Related MCP Connectors
Hosted MCP for YouTube Studio: uploads, metadata, playlists, comments, analytics, captions.
YouTube MCP — wraps the YouTube Data API v3 (BYO API key)
OAuth-protected, read-only-by-default MCP server for provenance-labeled QuillCaddie project memory.
YouTube discovery, transcripts, library search, and monitors with API keys or OAuth.
Related MCP Servers
- AlicenseBqualityDmaintenanceLocal read-only MCP server for YouTube Data API v3 and YouTube Analytics API, enabling public video searches and OAuth-based private channel reads.9MIT
- AlicenseAqualityDmaintenanceA local-first MCP server for inspecting and managing a creator-owned YouTube channel, using Google's official YouTube Data API v3 and Analytics API.1019 npmMIT
- FlicenseNot gradedqualityCmaintenanceMCP server for managing YouTube channels using Data API v3 and Analytics. Supports video upload, comments, playlists, and analytics via ~30 tools, with stateless OAuth authentication.-
- AlicenseNot gradedqualityAmaintenanceA local, privacy-first MCP server providing direct access to official YouTube Data, Analytics, Reporting, and Live Streaming APIs for creators, enabling channel analysis, research, and guarded management operations.3MIT