mcpField
F
FynePDF

FynePDF

Remote
fynepdf.comProductivity1Updated 6d ago

About FynePDF

FynePDF is a remote MCP server for assistants and agents that need to process PDF and related files. It offers document editing, conversion, OCR, signing, and AI-assisted reading, with a local bridge for transferring files between a computer and the service.

Key features

  1. 01Compress, merge, split, and edit PDFs
  2. 02Convert documents and create PDFs
  3. 03OCR scans and redact text
  4. 04Sign and protect PDF files
  5. 05Ask questions and extract document data
  6. 06Local bridge for file upload and download

Use cases

  1. 01Reduce PDF size or combine pages
  2. 02Redact sensitive text in documents
  3. 03Convert invoices into structured data
  4. 04Summarize PDFs with page citations

MCP support

The remote Streamable HTTP MCP server exposes file management, PDF operations, conversion, OCR, signing, and document analysis tools. The `npx @fynepdf/mcp` local bridge adds tools to upload local files and save results to disk.

Configuration

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

Endpointhttps://api.fynepdf.com/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.fynepdf.com/mcpAuth required
Packages
npm@fynepdf/mcp0.1.0

Tools

Details

From the project README

FynePDF

FynePDF MCP server

Real PDF tools for Claude, ChatGPT, Cursor, VS Code and any other MCP client:
compress, merge, split, convert, OCR, redact, sign, create PDFs, and read them with AI.

https://api.fynepdf.com/mcp

FynePDF is a remote MCP server (Streamable HTTP). Add the URL to an assistant once, sign in with your FynePDF account, and it can work on PDFs itself: "compress this to under 5 MB", "redact every phone number", "turn these invoices into a spreadsheet".

This repository holds the server's registry entry, the local bridge (npx @fynepdf/mcp) and a Claude skill. The server itself runs at FynePDF; docs are at fynepdf.com/developers/docs/mcp.

Connect it

AssistantHow
Claude (web, desktop, mobile)Settings, Connectors, Add custom connector, paste the URL. Guide
ChatGPTTurn on Developer mode in Settings, Apps, then create an app with the URL. Guide
CursorAdd to Cursor
VS CodeAdd to VS Code
Claude Codeclaude mcp add --transport http fynepdf https://api.fynepdf.com/mcp
Gemini CLIgemini extensions install https://github.com/fynepdf/fynepdf-mcp

Any client that speaks remote MCP works the same way: give it the URL, and it signs in with OAuth the first time a tool needs your account. Work then counts against your fynepdf.com plan, as on the website. Agents of your own can send an API key as the bearer token instead; that spends API credits.

Local files: npx @fynepdf/mcp

A remote server can't see files on your computer. The bridge runs on your machine over stdio, passes everything through to the remote server with your API key, and adds two tools of its own:

  • upload_local_file: sends a file on this computer to FynePDF and returns its file_id.
  • download_file: saves a result to your Downloads folder (or a path you name). It never replaces an existing file unless told to.

Name the file by its path ("compress ~/Downloads/report.pdf"). A file attached to the chat goes to the assistant's own sandbox, not your computer, so the bridge can't read it.

{
  "mcpServers": {
    "fynepdf": {
      "command": "npx",
      "args": ["-y", "@fynepdf/mcp"],
      "env": { "FYNEPDF_API_KEY": "fyne_live_..." }
    }
  }
}

Get a key in the developer dashboard: test keys are free (results are watermarked), live keys get 250 free credits a month. More in packages/mcp.

Tools

ToolWhat it does
upload_fileBring a file in: a drop zone in Claude, ChatGPT and VS Code, or a signed upload link
import_file_from_urlFetch a PDF, Office file, image or EPUB from a public link
get_fileA file's details and a fresh download link
list_filesRecent files (last 24 hours), to find one the person just uploaded by name
list_library_files, save_to_libraryUse and keep files in your FynePDF Library
merge_pdfs, split_pdf, edit_pagesJoin, split, delete, keep, reorder and rotate pages
stamp_pdfPage numbers, a watermark, or both
compress_pdfSmaller at a level, or under a target size
protect_pdfLock with a password, unlock, or flatten
redact_pdfRemove text for real: exact text, or described ("all names and emails")
convert_from_pdfTo Word, Excel or CSV, images, Markdown, EPUB, PDF/A, LaTeX or grayscale
convert_to_pdfFrom Word, Excel, PowerPoint, images, EPUB or a web page
ocr_pdfMake a scan searchable
create_pdfA report, invoice, letter or resume from Markdown, HTML or data, with the figures checked
sign_pdfStamp a signature image on a page
ask_pdf, summarize_pdfAnswers and summaries citing pages, with quotes checked against the text
extract_dataInvoices, statements and tables as data, added up in code
smart_editDescribe the change; FynePDF plans the steps and runs them
get_job, search_docs, list_actionsLong jobs, the docs, and every action with its price

Every tool returns a file_id, so the next tool works on the result without downloading and uploading again. In Claude and ChatGPT a result also shows as a card with a preview, Download and Save to Library.

Files and privacy

Files live on FynePDF only while they're worked on: they're deleted after 24 hours unless you save them to your Library, and download links last 1 hour. Disconnect an assistant any time in the dashboard; it stops working at once. See the privacy policy.

Also here

Prefer code? The REST API has the same actions, with SDKs for Node and Python.

License

MIT

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.