mcpField
F
flatmark

flatmark

Remote
flatmark.devProductivity0Updated 4w ago

About flatmark

flatmark converts PDF, Word, PowerPoint, Excel, and HTML documents to Markdown through an API and hosted MCP server. It offers one-call conversion for smaller files and queued OCR processing for larger or scanned documents.

Key features

  1. 01Convert documents to Markdown
  2. 02Support PDF, Word, PowerPoint, Excel, HTML
  3. 03Queued OCR for large files
  4. 04Table detection in queued jobs
  5. 05Return Markdown and JSON structure

Use cases

  1. 01Convert PDFs for text workflows
  2. 02Extract text from scanned documents
  3. 03Process large documents asynchronously

MCP support

The hosted MCP server is available at `https://api.flatmark.dev/mcp/` over Streamable HTTP. MCP clients can use its document conversion capabilities; anonymous access is supported with lower limits.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://api.flatmark.dev/mcp/
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

All endpoints
Streamable HTTPhttps://api.flatmark.dev/mcp/Auth required

Tools

Details

From the project README

flatmark

smithery badge

Document to Markdown API and MCP server for PDF, Word, PowerPoint, Excel and HTML. OCR queue for large files. Hosted in Germany.

flatmark converts PDF, Word, PowerPoint, Excel and HTML to Markdown over a REST API and an MCP server. Files up to 8 MB convert in one call with MarkItDown. Files up to 25 MB and 200 pages go through a queue that runs Docling with OCR and table detection. The queue returns Markdown and a JSON structure file, by polling or a signed webhook. The servers are in Germany. The free plan has 100 credits a month and needs no card.

Quickstart

Convert a PDF to Markdown in one call. The anonymous tier needs no key; add -H "X-API-Key: $FLATMARK_API_KEY" for your plan's limits.

curl -sO https://flatmark.dev/app-static/samples/sample.pdf
curl -F [email protected] https://api.flatmark.dev/v1/convert
{"markdown":"Hello Docling smoke test\n\n","meta":{"filename":"sample.pdf"}}

Files over 8 MB and scanned documents go through the OCR queue instead: POST /v1/convert/jobs, then poll or receive a signed webhook.

Use it from

MCP server

Streamable-HTTP endpoint: https://api.flatmark.dev/mcp/ — the anonymous tier works without a key; an API key raises the limits.

Claude Code:

claude mcp add --transport http flatmark https://api.flatmark.dev/mcp/

Claude Code: /plugin marketplace add flatmark-dev/flatmark then /plugin install flatmark@flatmark (set FLATMARK_API_KEY for your key).

Cursor / Windsurf / Cline / Claude Desktop:

{
  "mcpServers": {
    "flatmark": {
      "url": "https://api.flatmark.dev/mcp/"
    }
  }
}

One-click installs for every client: https://flatmark.dev/connect

Integrations (n8n, Zapier, Make, Dify, SDKs and templates): https://github.com/flatmark-dev/flatmark-integrations

Templates

Ready-made workflows: https://flatmark.dev/templates

Links

API

Base URL https://api.flatmark.dev/ — authenticate with an X-API-Key header; anonymous calls work at a lower rate limit.

About the files here

  • .claude-plugin/ — Claude Code plugin, with this repo as its own marketplace.
  • .mcp.json — the MCP server the Claude Code plugin connects to.
  • gemini-extension.json — Gemini CLI extension manifest.
  • glama.json — Glama server claim (names the maintainer).
  • mcp.json — Agent Plugins 1.0 MCP server config.
  • plugin.json — Agent Plugins 1.0 plugin manifest.
  • server.json — the entry in the official MCP Registry.
  • skills/ — an Agent Skill (agentskills.io SKILL.md) for skill galleries.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.