Bring Markdown
to OpenAI Agents SDK

Connect your Obsidian Publish environment to your AI agent and construct an intelligent oracle that reads smoothly from your personal or corporate markdown knowledge base.

What you can do

• Vault Crawling — Programmatically fetch your entire published vault structure utilizing list_files and list_navigation to build contextual trees.
• Direct Note Access — Execute get_file to stream the complete raw markdown contents of any note directly into your chat workflow for fast summarization.
• Metadata Operations — Use get_metadata to retrieve frontmatter properties, tags, and internal link logic mapped by Obsidian.
• Site Auditing — Easily ping site_info to ensure connectivity and verify the deployment status of your target Obsidian publish endpoint.

How it works

1. Subscribe to this server
2. Enter your Obsidian Publish Site ID (and optional access token for private vaults)
3. Start fetching and analyzing your personal notes directly from Claude, Cursor, or any MCP client

Who is this for?

• Researchers & Writers — seamlessly command an AI to read past chapters, cross-reference characters, or verify published assertions without opening Obsidian.
• Corporate Teams — build a highly accessible chatbot that rapidly reads and summarizes your team's documented standard operating procedures (SOPs).
• Developers — fetch specific code snippets and technical architectures housed in your digital garden natively within the IDE.