Ar9av/obsidian-wiki

What it solves

It prevents the loss of knowledge gained during AI agent interactions. Instead of letting insights be buried in chat logs, this tool allows an AI agent to compile and maintain a personal, interconnected markdown-based knowledge base (a "second brain") that the user owns and can read in Obsidian.

How it works

The system provides a set of "skills" (slash commands) that AI agents (like Claude Code, Cursor, or Windsurf) can use to ingest data from various sources—such as PDFs, chat exports, and URLs—and distill it into a vault of markdown files. It uses a manifest to track changes so it only processes new or updated information. It avoids simple accumulation by merging new knowledge into existing pages and flagging contradictions.

Who it’s for

Developers and power users who use AI agents to solve problems and want a persistent, structured, and searchable memory for their agents across different projects and sessions.

Highlights

• Agent-Driven Growth: AI agents can directly manage the wiki using built-in skills to capture, update, and query knowledge.
• Local Ownership: Knowledge is stored as plain markdown files in a local folder, ensuring no vendor lock-in.
• Graph-Based Intelligence: Improves the agent's ability to answer structural questions about how topics are connected, significantly reducing answer time and increasing accuracy compared to plain agents.
• Knowledge Validation: Tags claims as extracted, inferred, or ambiguous to distinguish between fact and speculation.
• Detailed Analytics: Supports exporting the knowledge graph to formats like GraphML, Neo4j Cypher, or interactive HTML.