Summary This project offers a local, AI-driven knowledge management system for non-technical users, integrating Obsidian with agents like OpenCode, Claude Code, and Codex. It automates ingestion, organization, querying, and maintenance of personal knowledge from diverse sources, creating a structured, accessible personal knowledge base.

How It Works The system combines Obsidian (local notes), an AI Agent (OpenCode, Claude Code, Codex), and Knowledge Base Rules (AGENTS.md). Users ingest content; the AI agent processes it per rules, structuring notes in Obsidian's wiki/ directory. This enables conversational querying and automated knowledge base health checks. Advanced features like social media digestion and web scraping are powered by tools like OpenCLI.

Quick Start & Requirements Installation uses a setup.sh script, primarily for macOS (12+). Linux users can adapt it; Windows users should use WSL2. An AI-assisted deployment option is also available. Prerequisites include Node.js. Estimated setup time is ~5 minutes. Detailed deployment guides are provided.

Highlighted Details

• Automated Ingestion: Processes articles, PDFs, screenshots into structured notes.
• Social Media Integration: Uses OpenCLI to scrape and digest content from platforms like Xiaohongshu, Douyin, Twitter, and Weibo, classifying it into knowledge domains.
• Conversational Querying: Allows natural language questions about stored knowledge.
• Knowledge Base Health Checks: AI performs regular checks for dead links, duplicates, and orphaned pages.
• Local-First Data Storage: Notes are local; AI processing sends data to configured model providers.
• Flexible AI Providers: Supports multiple agents and configurations for six external AI service providers.

Maintenance & Community Updates are managed via scripts/upgrade.sh for rules/skills. Uninstall scripts are available. Community interaction occurs via GitHub Issues and Discussions.

Licensing & Compatibility Released under the MIT License, permitting free use, modification, and distribution with attribution. Suitable for commercial use, but social media scraping may violate platform ToS.

Limitations & Caveats The primary install script targets macOS; Linux/Windows may need manual adjustments. While notes are local, AI-processed content is sent to third-party providers. Social media scraping is for personal learning/research only, with users responsible for ToS compliance and potential account risks.