About markymark
The problem
Section titled “The problem”If you work with Markdown files — documentation, notes, knowledge bases — you’ve probably experienced the pain of broken links, lost references, and no way to search across files. Renaming a heading means hunting through every document that links to it. Finding all references to a topic means grepping and hoping you catch everything.
What markymark does
Section titled “What markymark does”markymark gives your text files the same navigation and refactoring tools that programmers expect for source code. Open a workspace and you get:
- Go to definition — jump to any heading, wiki link, or block ID across your files
- Find all references — see every document that links to a heading or symbol
- Rename everywhere — rename a heading and update all references in one step
- Broken link detection — see dead links and duplicate anchors as you type, before they become a problem
- Workspace search — fuzzy symbol search, full-text search, and regex with context lines
- Multi-format support — beyond Markdown, markymark understands JSON, JSONC, JSON5,
JSONL, YAML, TOML,
.env, and INI files - Obsidian and Logseq support — wiki links (
[[page]]), callouts, block IDs, and page properties work out of the box
How it works
Section titled “How it works”markymark runs in two modes:
Inside your editor. markymark integrates with editors like VS Code and Neovim through a protocol called LSP (Language Server Protocol). You install an extension, point it at a folder, and your editor gains all the navigation and diagnostic features listed above — with results appearing as you type.
As an AI agent tool. AI coding assistants like Claude Code and Cursor can connect to markymark as a tool server (using a protocol called MCP). This lets agents read, search, and navigate your documents programmatically — the same capabilities a human gets in their editor, but accessible to automated workflows.
Under the hood, markymark is written in Rust with a dependency-free Zig parser core (based on Bun’s Zig port of md4c), designed for fast response times even on large workspaces.
Who it’s for
Section titled “Who it’s for”Developers maintaining docs-as-code. If your documentation lives alongside your code in Markdown files, markymark catches broken links, helps you refactor headings safely, and gives you workspace-wide search without leaving your editor.
Knowledge workers with large note collections. If you use Obsidian, Logseq, or any tool that stores notes as Markdown files, markymark adds IDE-level navigation: jump between notes, find everything linking to a page, and spot broken references across hundreds of files.
AI agents that need document access. If you’re building workflows where AI agents read and reason about documentation, markymark provides structured access to document content, symbols, and cross-references — without the agent needing to parse raw files itself.
Current status
Section titled “Current status”markymark is in active pre-release development. The core feature set is stable and used daily, but APIs and behavior may change before v1.0. The project is licensed under AGPL-3.0.
Source code and releases: github.com/sethyanow/markymark