PDFGate MCP Server

What PDFGate MCP Server MCP server does

The PDFGate MCP Server MCP server connects an MCP client to PDFGate’s document and e-signature API. It can create PDFs from a URL or raw HTML, upload existing PDFs, retrieve document metadata, download file bytes, and remove stored documents. Operations that modify a document can create a new stored document while preserving the source, including flattening, protection, compression, and watermarking.

Form workflows are supported in both directions. The server can read values from fillable PDFs and add interactive fields using placeholder tags, field overrides, or explicitly positioned fields. Watermarks can use text or an image, while password protection can include permission restrictions.

Signing workflows use envelopes built from one or more stored documents. Agents can create an envelope, send it to recipients, inspect signing progress, void active envelopes, or delete eligible envelopes and their produced files. Embedded recipients can sign inside an application through a short-lived link created for an iframe session.

How it works

PDFGate documents are referenced by IDs returned from upload or generation operations. Those IDs are then supplied to processing, download, and envelope tools. Envelopes begin in a created state and must be sent before PDFGate dispatches signing emails. Embedded recipients do not receive email; an application creates a signing link after the envelope is in progress.

Recipient records can be stored and reused by ID. The server also exposes webhook management. Registered events include envelope sent, completed, expired, voided, deleted, document completion, and recipient signing. Incoming events are available through the pdfgate://events MCP resource.

Setup and configuration

Run the PDFGate MCP Server MCP server locally through the @pdfgate/mcp-server npm package and provide PDFGATE_API_KEY. Keys beginning with live_ select production, while test_ keys select the sandbox. The webhook listener uses port 3599 by default and can be changed with PDFGATE_WEBHOOK_PORT.

A hosted option is available at https://mcp.pdfgate.com/server. MCP clients that support remote Streamable HTTP servers can authenticate through OAuth, or send an X-API-KEY header. OAuth sessions use the production account; sandbox access uses a test_ key. To receive webhook events, expose the listener to the internet and register the public URL with create_webhook.

Tools and capabilities

• Generate PDFs from URLs or HTML and upload local files or remote URLs.
• Retrieve, download, delete, flatten, compress, protect, and watermark documents.
• Extract fillable form data and add interactive fields.
• Create, send, inspect, void, and delete signing envelopes.
• Create embedded signing links and manage stored recipients.
• Create, inspect, and delete webhook subscriptions.

Limitations and notes

Generation accepts either a URL or raw HTML, not both. Uploading likewise accepts either a local file path or a URL. Flattening and protection produce new documents rather than overwriting the original. Deleting an envelope also removes its signed documents and audit logs, cannot be undone, and is restricted by envelope status; active envelopes must be voided first.

Embedded links expire after 10 minutes and should be created when the signer is ready. A separate link is needed for each signing session. The local webhook listener requires an internet-accessible endpoint, such as one exposed through a tunneling service. The provided material does not specify a software license or PDFGate pricing.