Best MCP servers for PDFs and documents
10 working options, checked Nov 8, 2026.
Short answer: the top picks for PDFs and documents right now are ExtractPDF, Agent Extraction API (FreeTier Agent API) and CleverUtils MCP Server. We list 10 options here: 7 work without sign-in, 1 need a sign-in and 2 run on your own machine.
Document MCP servers let an AI assistant pull text and data out of PDFs, Word files and scans, convert between formats, or publish a document as a shareable link. They save you from pasting long files into a chat by hand.
Top picks
-
ExtractPDF
Extract text from any PDF or DOCX — $0.01 USDC per call, no API key needed
Tools include
extract_document -
Agent Extraction API (FreeTier Agent API)
Pay-per-call API and MCP server for AI agents: invoice PDF to JSON, clean web pages, contact extraction.
Tools include
list_endpoints,mcp_servers_sample,llm_pricing_sample,api_directory_sample,invoice_json -
CleverUtils MCP Server
Remote MCP server for CleverUtils, a free file converter. Lets AI agents convert audio, video, images, PDFs, documents, ebooks and 3D files across 200+ converters, with no API key.
Tools include
convert_file,upscale_image,remove_background,colorize_photo,restore_old_photo -
OCR-AI
Hebrew OCR: documents to text, fields and Excel
Tools include
read_document,check_balance,buy_pages -
AgentRender
Screenshot, PDF, and structured extract for any public URL
All 10 compared
| Server | What it does | Setup | Tools |
|---|---|---|---|
| ExtractPDF | Extract text from any PDF or DOCX — $0.01 USDC per call, no API key needed | Hosted, no sign-in | 1 |
| Agent Extraction API (FreeTier Agent API) | Pay-per-call API and MCP server for AI agents: invoice PDF to JSON, clean web pages, contact extraction. | Hosted, no sign-in | 74 |
| CleverUtils MCP Server | Remote MCP server for CleverUtils, a free file converter. Lets AI agents convert audio, video, images, PDFs, documents, ebooks and 3D files across 200+ converters, with no API key. | Hosted, no sign-in | 15 |
| OCR-AI | Hebrew OCR: documents to text, fields and Excel | Hosted, no sign-in | 3 |
| AgentRender | Screenshot, PDF, and structured extract for any public URL | Runs locally | – |
| switch402 | Web scraping, search, screenshots and PDF to text for agents, paid per call over x402. No signup. | Hosted, no sign-in | 9 |
| StatementKit | Normalize bank statements (CSV, camt, MT940, OFX, QIF) to JSON, analyze and export them. | Hosted, no sign-in | 5 |
| Filesystem | Read, write, and search files in allowed directories on the host machine. | Runs locally | – |
| dochost | Let your AI assistant publish HTML or Markdown as a shareable web link. | Hosted, no sign-in | 8 |
| HaveTML | Publish the HTML, Markdown, JSON or YAML your agent makes as a permanent sandboxed share link. | Hosted, sign-in | – |
How to choose
- Scanned or digital? Scans and photos need OCR. Files with real text inside are simpler and cheaper to read.
- Plain text or structured fields? For invoices and forms, choose a server that returns fields as JSON.
- Mind confidential files. Hosted servers receive your document. Use a local option for sensitive material.
How to connect one
Here is ExtractPDF as an example. In Claude Code, run:
claude mcp add --transport http extractpdf https://www.extractpdf.xyz/mcpFor other apps, follow our guides for Claude, ChatGPT, Cursor and VS Code.
How we picked these
We read each listing's tools and put the closest fit for this job first. Hosted servers must pass a real MCP handshake in our daily check to stay on this page, and local servers are reviewed from their source or package. That shows a server responds, not how well it works for you, and we don't run every tool in production.
Frequently asked questions
Which of these MCP servers work without signing in?
ExtractPDF, Agent Extraction API (FreeTier Agent API), CleverUtils MCP Server, OCR-AI and switch402 are among the 7 that answered our latest live check with no account or key.
How do I let Claude read a PDF through MCP?
Connect a document server, then give the assistant a file or URL. The server extracts the text and returns it for summarizing or searching.
Can MCP do OCR on scanned documents?
Yes. OCR servers convert images of text into real text, and some handle specific languages or return structured fields.
How often is this list updated?
Every hosted server here is re-checked daily with a real MCP handshake, and a server drops off this page while it fails that check. Last check: Nov 8, 2026.