YangLiangwei/PersonalizationMCP
PersonalizationMCP is a comprehensive Model Context Protocol (MCP) server that aggregates personal data from multiple online platforms. Built in Python, it integrates with services like Steam, YouTube, Bilibili, Spotify, and Reddit, allowing AI assistants (such as Claude or any MCP-compatible client) to access user data through a standardized interface. The server handles OAuth2 authentication, automatically manages access tokens and refresh flows, and provides a unified API for querying personalized content. The tool includes an interactive CLI onboarding process that guides users through setting up API credentials for each platform. It features a modular architecture with separate platform integrations and a shared token store, making it easy to enable only the services you need. The server exposes MCP tools and resources that map to platform-specific data, such as game libraries, subscription lists, playlists, saved posts, and more. Key capabilities include retrieving Steam game libraries with playtime and achievements, listing YouTube subscriptions and playlists, fetching Bilibili video history and favorites, accessing Spotify listening data and playlists, and pulling Reddit user activity (posts, comments, saved items). The server is designed to run locally or on a personal server, keeping data under the user's control while extending AI assistants with rich personal context.
- Category
- Developer Tools
- Pricing
- Open Source
Problem solved
Users often have valuable personal data scattered across multiple platforms, but AI assistants cannot natively access this information due to API complexity and authentication barriers. PersonalizationMCP solves this by acting as a bridge: it securely authenticates via OAuth2 to each service, aggregates the data, and exposes it through the MCP standard. This enables AI tools to answer questions like 'What games do I own?' or 'Summarize my saved Reddit posts' without the user manually exporting data.
Use cases
- Ask an AI assistant to list all games in your Steam library and recommend which to play next based on playtime and achievements.
- Retrieve and summarize your latest YouTube subscription videos directly in a conversation.
- Fetch your saved Reddit posts and have an AI compile them into a research document.
- Query your Spotify listening history to generate a monthly music recap.
- Combine data from Bilibili favorites and YouTube playlists to create a unified media watchlist.
Audience
- Developers
- Hobbyists
- ML Engineers
- Data Scientists
Pros
- Multi-platform support: aggregates Steam, YouTube, Bilibili, Spotify, Reddit, and more.
- Secure OAuth2 authentication with automatic token management and refresh.
- MCP-compliant, works with any MCP client (Claude Desktop, Continue, etc.) out of the box.
- Interactive CLI onboarding simplifies the setup of API credentials.
- Modular architecture allows enabling only desired platforms, reducing overhead.
Cons
- Requires users to obtain and configure API keys from each platform (no zero-config setup).
- Limited to the implemented platforms; adding new services requires custom development.
- Potential privacy risk as personal data is exposed to the connected AI model, so careful permission scoping is advised.
Official website: https://github.com/YangLiangwei/PersonalizationMCP
Canonical LambdaBase page: https://www.lambdabase.com/tools/yangliangwei-personalizationmcp