---
title: "mdprint developer documentation: local WebMCP, public API & agent skills"
url: "https://mdprint.app/developers/"
dateModified: "2026-09-21T18:40:36+02:00"
---

# mdprint for developers & agents

mdprint.app is a free browser tool by Fokuus BV for turning Markdown into a printable document or PDF. Use an agent to prepare the Markdown, then open it in mdprint to review and export.

## Use the browser workflow

The browser editor remains the conversion interface: it processes documents on your device without an account. WebMCP tools expose the editor to supported browser agents. A separate public API and MCP server provide read-only documentation and examples; they do not convert or accept private documents.

1.  Prepare a Markdown document with headings, lists, tables and fenced code blocks. Save it as `.md`, `.markdown` or `.txt`.
2.  Open [the mdprint editor](https://mdprint.app/#tool). Paste the text, use Import, or drop the file into the editor.
3.  Review the preview and choose the page size, margins and orientation in Settings. Check tables, code blocks, math and Mermaid diagrams before exporting.
4.  Choose Print or Download PDF using the guidance below. Inspect the resulting pages before sharing them.

An agent with browser access can assist with this workflow using the visible controls. File access, downloads and the native print dialog depend on that agent’s browser environment; the user may need to finish the export.

## Local WebMCP and public documentation API

### Browser tools: conversion stays on your device

In browsers supporting WebMCP, the editor registers three tools: `mdprint_get_document_info`, `mdprint_set_markdown` and `mdprint_export_pdf`. The first returns metadata and a revision, not your private document. Editing and export require the current revision; replacing existing work also requires an explicit replace flag. A pending recovered draft must be resolved first.

WebMCP is experimental and browser support varies. The regular controls remain available. The PDF download is saved through the browser, not returned to a remote server. An agent running its browser in the cloud uses that cloud machine, rather than your laptop.

### Public read-only API

No signup, API key or payment is required. [Product information](https://mdprint.app/api/v1/info), [Markdown examples](https://mdprint.app/api/v1/examples) and [documentation sections](https://mdprint.app/api/v1/docs) return JSON. Add `?id=basic` to examples, or `?id=privacy` to documentation, to read one item. Invalid parameters return a JSON error. Requests are rate-limited.

Read the [OpenAPI 3.1 specification](https://mdprint.app/openapi.json) and [authentication and access guide](https://mdprint.app/auth.md). These interfaces accept no user document and do not call the private conversion prototype.

```
curl https://mdprint.app/api/v1/examples?id=basic
```

### JavaScript SDK and command-line client

The official `@fokuus/mdprint-client` SDK and `mdprint-docs` CLI read the public documentation and examples. Node.js 20 or newer is required. Install the website-hosted release directly:

```
npm install https://mdprint.app/developers/downloads/fokuus-mdprint-client-0.1.1.tgz
npx mdprint-docs info
npx mdprint-docs examples tables
```

```
import { createClient } from '@fokuus/mdprint-client';
const mdprint = createClient();
const info = await mdprint.getInfo();
```

Read the [SDK documentation and public wrapper source](https://github.com/Fokuus/ai-agent-templates/tree/main/integrations/mdprint-client), or [download the package](https://mdprint.app/developers/downloads/fokuus-mdprint-client-0.1.1.tgz). This release is hosted on mdprint.app; an npm registry listing is not required. The client makes read-only requests and has no conversion engine or runtime dependencies.

### Search public documentation with NLWeb

The [documentation search endpoint](https://mdprint.app/ask?query=PDF) supports NLWeb 0.55 list mode and optional server-sent events. It uses keyword matching over a small fixed collection of public documentation and examples. It returns source links, not generated answers. Use short search queries; do not send documents or conversation history.

```
curl 'https://mdprint.app/ask?query=PDF&mode=list&streaming=true'
```

POST JSON uses a `query` object with `text`, plus an optional `prefer` object with `mode: list` and `streaming: true`. Search queries are limited to 300 characters. Empty matches return `NO_RESULTS`; unsupported answer-generation modes return `UNSUPPORTED_MODE`.

### Documentation MCP server

Connect an MCP client using Streamable HTTP to `https://mdprint.app/mcp/docs`. Tools: `get_product_info`, `list_examples`, `get_example` and `get_docs`. They provide the same public information as the API, without authentication or document uploads. Discover connection metadata in the [server card](https://mdprint.app/.well-known/mcp/server-card.json).

```
{
  "servers": {
    "mdprint": {
      "type": "http",
      "url": "https://mdprint.app/mcp/docs"
    }
  }
}
```

The configuration above can be used in VS Code's MCP settings. It gives an agent documentation access; file conversion still uses your browser. Client software controls connection trust and tool permissions.

## Choose the right PDF export

### Print → Save as PDF

Opens the browser’s print dialog. Choose Save as PDF to keep document text selectable and searchable. This is the preferred route when the PDF needs to be read by people and text extraction tools. Review the browser’s paper size, margins and header/footer settings.

### Download PDF

Creates a PDF directly in the browser using the app’s page settings. Pages are rendered as images, so text in this export is not selectable or searchable. Use it for a visual copy; keep the original Markdown for editing and agent workflows.

## Local files, privacy and limitations

Markdown rendering, file imports and PDF generation run on your device. Document content is not uploaded to a conversion server. Page views and aggregate action counts are measured separately; see the [privacy policy](https://mdprint.app/privacy/).

External image URLs may be blocked by the site’s security policy; do not assume remote images will appear in the export. Loading the website requires network requests for application assets, fonts and analytics. Local document processing does not mean every browser request is offline.

The [PDF to Markdown tool](https://mdprint.app/pdf-to-markdown/) extracts selectable text locally. It does not perform OCR on scanned pages. PDF layout does not preserve the original Markdown structure, so complex tables, columns and reading order can need manual correction. An image-based PDF from Download PDF cannot be converted back into text with this tool.

## Official templates and agent resources

These public repositories are maintained by Fokuus. They contain templates and agent instructions for working with mdprint.

*   [Markdown templates](https://github.com/Fokuus/markdown-templates): starting documents to adapt and open in mdprint.
*   [AI agent templates](https://github.com/Fokuus/ai-agent-templates): reusable instructions, including the mdprint skill.
*   [Read the mdprint skill](https://mdprint.app/skills/mdprint/SKILL.md) for an agent-ready workflow and export limitations.
*   [llms.txt](https://mdprint.app/llms.txt): a concise product description and resource links.
*   [Agentic Resource Discovery catalog](https://mdprint.app/.well-known/ard.json): machine-readable resources available from mdprint.

### Install the agent skill

In a terminal with Node.js and npm available, use the Skills CLI to add the instructions to your supported coding agent:

```
npx skills add Fokuus/ai-agent-templates --skill mdprint
```

The skill teaches the local browser workflow and links the public documentation interfaces. It does not install a PDF conversion service or grant file access.

## Ready to print your Markdown?

Open the editor to preview your Markdown and choose your PDF export.

[Open the editor](https://mdprint.app/#tool) [Read the guide](https://mdprint.app/how-to-use)

---
Canonical page: https://mdprint.app/developers/

[Agent documentation](https://mdprint.app/developers/) · [Authentication and privacy](https://mdprint.app/auth.md) · [Resource catalog](https://mdprint.app/.well-known/ard.json)