Write, share, and carry your context everywhere. Create your free account today.

01 Welcome to MDflow

Raw

Welcome to MDflow

MDflow is an online markdown workspace for people who want to write in a portable format, share clean documents with readers, and make the same knowledge available to scripts and AI agents.

This documentation covers everything you can do in MDflow — the web app, the companion apps, and the developer and AI interfaces.

What MDflow is

MDflow is three things at once:

  • A markdown editor in your browser. There is no desktop application to install. You write in GitHub Flavored Markdown, see a live rendered preview, and keep everything as plain text you can take with you at any time.
  • A place to organize and share knowledge. Workspaces, nested folders, collections, full-text search, and share links that readers open without needing an account.
  • A shared brain for you and your AI agents. The documents you write are readable and writable through an HTTP API and Model Context Protocol (MCP) servers, so Claude, ChatGPT, Cursor and your own scripts work on exactly the same content you see on screen.

Why plain markdown

Markdown is just text. That has three practical consequences:

  • Nothing is locked in. You can download any document as a .md file, or export an entire workspace as a .zip archive that renders offline in Obsidian, VS Code, or on GitHub.
  • It travels well. The same document reads correctly in the browser, in your code editor, on your phone, and in a language model's context window.
  • It diffs cleanly. Because documents are text, MDflow can show you a line-by-line comparison between any two versions.

What you can do

AreaWhat it covers
WriteA Monaco-based editor with Edit, Split and Preview views, live preview, focus mode, images, internal links, autosave and version history
OrganizeWorkspaces, nested folders with descriptions, collections, favorites, full-text search, and a Tasks view built from markdown checkboxes
SharePublic read-only links, private sharing by email, shared collections with an optional searchable Knowledgebase mode, and comments
ProtectPer-document encryption in the browser, where the server never sees your text or your password
ConnectGitHub and Google Drive integrations, a Web Clipper extension, a VS Code and Cursor extension, an iOS app, and an n8n node
AutomatePersonal Access Tokens, a versioned HTTP API, hosted and local MCP servers, and machine-readable agent discovery files

A note on structure

MDflow has three levels of organization:

  1. Workspaces are the top level. Keep unrelated areas of life apart — Work, Personal, a side project. Switching workspaces swaps the whole sidebar.
  2. Folders live inside a workspace and can nest as deeply as you like. Each folder has a description, which is both a note to your colleagues and the primary signal AI agents use to find relevant content.
  3. Documents are the markdown files themselves.

Collections sit outside this hierarchy on purpose. A collection gathers documents from across any of your workspaces into a named reading set, without moving or copying anything.

Where to go next

MDflow is at https://mdflow.cz.