> Automate MDflow with n8n.

The action node exposes the whole MDflow object model as ordinary n8n operations — pick a resource, then an operation, and map your workflow data onto the fields.

The MDflow Trigger node polls for new content, so a workflow can fire the instant something is created:

• ›New Document — watches a folder.
• ›New Folder — watches an optional workspace, or the whole account.
• ›New Workspace — watches the whole account.

• ›500+ integrations, no code. Everything n8n already connects to — forms, databases, GitHub, CRMs, Slack, email, RSS, webhooks — can now read from and write to your markdown workspace.
• ›Two-way, not just a sink. The node has full read and write access: pull a document's body into a workflow, transform it, and write it straight back.
• ›Event-driven. The trigger node turns “a new document landed” into the start of an automation — index it, back it up, notify a channel, kick off a review.
• ›A workspace your AI agents use. The MDflow node works as a tool for n8n AI Agents, so an agent can manage the same markdown your team edits on the web.
• ›One account, every surface. The documents you automate here are the same ones you reach on the web, on mobile, in the VS Code extension, and over the MCP server.

• ›Capture forms as markdown. An n8n form or a webhook submission becomes a formatted document in the right folder — the exact flow shown above.
• ›Build a knowledge base on a schedule. Pull from GitHub issues, a database, RSS, or a CRM, render the result as markdown, and create or update documents so your workspace stays an up-to-date, agent-readable source of truth.
• ›Publish and announce automatically. Create a document, use Set Sharing to publish it at a link, then post that link to Slack, Discord, or email in the same run.
• ›React to new content. Let MDflow Trigger fire on a new document to back it up, index it elsewhere, or notify a reviewer.
• ›Send scheduled digests. A cron trigger gathers the day's data, writes a markdown report into MDflow, and shares the rendered page — a standing daily brief with no manual step.

MDflow is built for people and AI agents, and n8n is where the two meet. There are two ways to put your markdown in front of an agent:

• ›As an AI Agent tool. Connect the MDflow node to an n8n AI Agent node's tool input and the agent can create, read, update, and share documents as part of its reasoning.
• ›Over the remote MCP server. For free-form access, point n8n's MCP Client Tool node at MDflow's hosted server — https://mdflow.cz/api/mcp — and the agent gets the full MCP toolset, including mdflow_get_context retrieval ranked by folder descriptions.

• ›Update Body replaces the whole body. To append instead, Get the document first and concatenate in an expression before the Update Body operation.
• ›Size limits. Document bodies are capped at 500 KiB and folder descriptions at 5,000 characters — the same limits as everywhere else in MDflow.
• ›Duplicate names are auto-suffixed. The server disambiguates duplicate document titles and folder names with a numeric suffix and returns the final name.
• ›Triggers poll. The trigger node checks for new content on n8n's schedule rather than receiving a push, so expect it to fire within a polling interval, not instantly.