whiteguo233/OpenBiliClaw

What it solves

OpenBiliClaw is a personalized content discovery agent that breaks the "information cocoon" created by platform-controlled recommendation algorithms. It solves the problem of fragmented user interests across different platforms (like Bilibili, YouTube, and X) and the lack of transparency and control over why certain content is recommended.

How it works

The system operates as a local AI agent that builds a deep "soul portrait" of the user through a five-layer memory system (events → preferences → awareness → insight → soul). Instead of simple tag matching, it uses LLMs to analyze user behavior and feedback to infer cognitive styles and psychological needs. It then actively searches across multiple platforms—including Bilibili, Xiaohongshu, Douyin, YouTube, X, Zhihu, Reddit, and Bangumi—to find content the user might like, even guessing new interests based on psychological bridges.

Who it’s for

Users who want a private, cross-platform content discovery experience where they own their data and can actively "train" their AI assistant to understand their tastes and intellectual curiosity without being manipulated by platform-driven engagement metrics.

Highlights

• Cross-Platform Integration: Aggregates content from a wide variety of social and video platforms into a single local interface.
• 100% Local & Private: All user data is stored in a local SQLite database, and LLMs are used via API keys or local providers like Ollama.
• Active Exploration: Unlike passive collaborative filtering, it actively hypothesizes and tests new interest areas to help users break out of filter bubbles.
• Detailed Persona Analysis: Generates natural language descriptions of the user's personality, MBTI, and core needs to explain recommendation logic.
• Multi-Device Access: Includes a browser extension for data collection and a web interface for desktop and mobile viewing.