# Markdown Frontmatter Harvester MCP for AI Agents AI Agent Connect > Markdown Frontmatter Harvester scans your local Obsidian, Hugo, or Logseq vaults to grab all YAML metadata, like tags, status, and dates, and turns it into a single JSON index. This lets your AI client instantly query your entire knowledge base without having to open and read all your files manually. ## Overview - **Category:** developer-tools - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_pSzsBgFkLuKZKUA1Ej60Dr18di9s0JV9d52i5cdI/ai-agent-connect - **Tags:** yaml-parsing, metadata-extraction, markdown, obsidian, vault-management, structured-data ## Description The Markdown Frontmatter Harvester turns your messy folder of notes into a searchable database for your AI agent. If you use Obsidian or Hugo, you probably already have notes scattered everywhere with tags and status updates hidden in the frontmatter. The problem is that your AI client usually can't see that stuff unless it opens all your files one by one, which is slow and often fails on big vaults. This Connector changes that by acting as a high-speed librarian. It crawls your folders, pulls out just the metadata, and builds a clean map of your information. You can then ask your AI to find all 'draft' notes from last month or list all research on a specific topic, and it happens instantly because the data is already organized. It's a much faster way to work with your personal knowledge base, and it's one of the many tools you can find in the Vinkius marketplace to make your AI more useful. Everything stays local, so your private journals and business notes don't leave your machine. ## Tools ### harvest_markdown_frontmatter Provide a folder path to scan all markdown files and extract metadata like tags and dates. This lets your AI see the structure of your entire vault at once. ## Prompt Examples **Prompt:** ``` Find all notes in my Obsidian vault that are marked as 'urgent' and have the 'research' tag. ``` **Response:** ``` I found 8 notes matching those criteria. Here they are: * **ProjectX_Notes.md** * **Meeting_Notes_04.md** * **Lab_Results_Final.md** * **Draft_Proposal_A.md** * **Site_Audit_Notes.md** * **Urgent_Fix_01.md** * **Review_Notes_03.md** * **Quick_Thoughts_Bio.md** ``` **Prompt:** ``` Which of my blog posts are still in 'draft' status? ``` **Response:** ``` Based on the frontmatter in your Hugo directory, you have **12 posts** currently marked as 'draft'. Would you like me to list the filenames for those specific posts? ``` **Prompt:** ``` How many notes did I create in 2024 based on the date field? ``` **Response:** ``` According to the metadata, you created exactly **145 notes** in the year 2024. ``` ## Capabilities ### Index entire vaults Turns scattered local markdown files into a unified database. ### Extract YAML metadata Pulls tags, dates, and status fields from frontmatter. ### Filter notes by status Let your agent find drafts or published posts in milliseconds. ### Query specific tags Instantly see every note associated with a specific research topic. ### Scan large folders Handles thousands of files without slowing down your workflow. ## Use Cases ### Finding drafts in a large blog A blogger has 200 posts in a Hugo folder and needs to find the 5 that are still marked as 'draft' to finish them today. ### Tag-based research discovery A researcher wants to see every note tagged with 'biotech' and '2023' to prepare for a meeting. ### Date-based auditing A student wants to see how many notes they wrote in October to track their study progress. ### Status tracking across folders A project manager wants to list all notes marked as 'urgent' across multiple different folders. ## Benefits - Fast searching: Use harvest_markdown_frontmatter to find notes by tag or date in milliseconds instead of waiting for your AI to read every file. - Better organization: Turn your scattered Obsidian notes into a structured database that your AI can actually navigate. - Local privacy: Your notes stay on your hard drive because the Connector processes everything locally without sending your files to the cloud. - Bulk status checks: Quickly identify every 'draft' or 'in-progress' note in a large Hugo repository using metadata extraction. - Instant filtering: Ask your agent to group notes by specific categories or dates without manual sorting. ## How It Works The bottom line is it turns your flat folder of files into a queryable database for your AI agent. 1. Provide the absolute path to your local Markdown folder. 2. The Connector scans your files and pulls out the YAML frontmatter. 3. You get a structured index that your AI client can query instantly. ## Frequently Asked Questions **Does the Markdown Frontmatter Harvester work with my Obsidian vault?** Yes, it's designed to scan your local Obsidian folders and pull out all the YAML metadata. **Is my data safe with the Markdown Frontmatter Harvester?** Your notes never leave your machine. The Connector scans your local files and only shares the metadata with your AI client. **Can I use this for my Hugo website?** Yes, it works great for Hugo sites by extracting the frontmatter from your blog posts. **How fast is the indexing?** It uses high-speed globbing to scan thousands of files in milliseconds. **Does it read the actual content of my notes?** No, it only looks at the YAML frontmatter at the top of your files. **Do I need to move my files to a special folder?** No, you just need to give the Connector the absolute path to your existing folder. **Does it send my entire note content to the AI?** No. To protect your privacy and save context tokens, it strictly extracts ONLY the YAML metadata block between the `---` delimiters. The actual body text of your note is ignored. **Will this alter my Obsidian files?** Not at all. The parser operates in a strict read-only mode. It reads the files into memory to extract the metadata and never writes back to the disk. **Does it work with nested folders?** Yes! The engine uses recursive globbing to scan every single folder and sub-folder inside the path you provide.