Enconvert

What the Enconvert MCP server does

The Enconvert MCP server exposes EnConvert’s file-conversion and web-data endpoints through the Model Context Protocol. It gives an agent a single interface for turning documents, images, URLs, and web search results into artifacts or structured outputs.

File tools handle document conversion between supported formats, image format changes, image compression, PDF creation, and Markdown extraction. Supported inputs include Office and OpenDocument files, iWork formats, EPUB, HTML, Markdown, CSV, JSON, XML, YAML, TOML, RTF, and several image formats. A job-status tool lets an agent check an individual asynchronous file-conversion job.

The web side can render pages into Markdown, HTML, screenshots, PDFs, links, and images. It also supports URL discovery, Google-backed search, schema-based extraction, RAG-oriented ingestion, and page watchers that retain change history.

How it works

The server presents 24 tools, seven prompts, and four read-only resources. The prompts package multi-step workflows such as converting a folder, building a site corpus, researching with cited sources, diagnosing a render, or creating a page monitor. Resources expose the conversion format matrix and information about ingest jobs, watchers, and watcher snapshots without triggering a browser render.

URL and browser operations use EnConvert’s V2 endpoints. A page can be rendered once, a batch can process up to 1,000 URLs, and structured extraction can work across up to 50 URLs or a discovered site. Large ingestion batches run asynchronously and can be polled or cancelled. Watchers can be created, updated, inspected, and deleted, with hourly-or-longer monitoring supported by the API.

Setup and configuration

The Enconvert MCP server requires Node.js ^20.20.0 || >=22.22.0 and an EnConvert API key. The documented setup command is npx @enconvert/mcp setup; its wizard detects supported clients, validates the secret key, and writes client configuration. Restart the selected clients after setup.

The key is stored in ~/.enconvert/config.json with restrictive permissions. ENCONVERT_API_KEY overrides the saved key and is suitable for manual configuration, Docker, or CI. The server can also be configured directly with npx -y @enconvert/mcp@latest and that environment variable. Management commands include status, rotate-key, upgrade, and remove.

Tools and capabilities

The Enconvert MCP server includes capabilities for:

• Converting documents or images between supported formats.
• Compressing PNG, JPEG, and WebP files while preserving the input format.
• Turning supported files into Markdown or PDF.
• Rendering URLs and retrieving signed artifact URLs.
• Discovering site URLs through sitemap, crawl, or hybrid methods.
• Searching web, news, image, scholar, patent, and map sources.
• Extracting structured data using a supplied schema.
• Creating and managing asynchronous RAG ingestion jobs.
• Creating page watchers and reviewing their snapshots and diffs.

For SVG conversion, optional width and height controls apply to SVG-to-PNG, JPEG, or WebP output. Image compression can accept a target size, but an unreachable target is not treated as an error; callers should inspect the returned file size.

Limitations and notes

The V2 web tools require a private EnConvert API key and are plan-gated. Webhook signing-secret endpoints are intentionally not exposed through MCP; those secrets must be retrieved from the EnConvert dashboard. Animated APNG and animated WebP inputs are rejected by image compression.

The setup wizard supports Claude Code, Claude Desktop, Cursor, Windsurf, VS Code, Zed, Gemini CLI, Codex CLI, and OpenCode. Native Windows manual configuration needs a cmd /c npx wrapper rather than invoking bare npx.