
pdf.net已验证已精选
@pdf.net
基本信息
分类
Web 聊天与 UI
作者
pdf.net
关于 pdf.net
Automate document workflows with pdf.net MCP server. Connect ChatGPT and Claude to generate, translate, merge, and edit PDFs with AI.
概览
Give your AI assistant a real PDF workspace. Upload, generate, edit, convert, merge, split, compress, search and organize documents in a persistent pdf.net library, straight from Claude, ChatGPT, Cursor or any MCP client.
pdf.net is a browser based PDF editor used by 2M+ customers. This server exposes that engine over MCP, so the assistant does not just read a file you paste into a chat. It keeps a real document library, edits text inside the PDF without a round trip through Word, produces real files, and hands you a link to open any of them in the pdf.net editor.
Hosted remote server. No install, no local runtime, no API key. Available on all pdf.net plans, including free.
Quick start
Endpoint: https://mcp.pdf.net/mcp (Streamable HTTP, OAuth 2.1)
Claude Code
claude mcp add --transport http pdf-net https://mcp.pdf.net/mcp
Claude (web and desktop)
Settings → Connectors, search for "pdf.net", click Connect, sign in to authorize. Available in every new conversation after that.
ChatGPT
Settings → Connected apps, find pdf.net, follow the same authorization flow.
What it does
Upload and import. Files from your device or from a URL. PDF, DOCX, DOC, TXT, XLSX, CSV, PPTX, PNG, JPG, GIF. Non PDF formats are converted automatically.
Edit documents. Change text, restyle it, add or remove pages, reorder or duplicate pages, move text blocks. Applied natively inside the PDF, keeping the original layout, fonts and formatting.
Generate documents. A complete formatted PDF from a plain language description, or a template filled in with specific details.
Convert. Export to docx, doc, odt, rtf, txt, html, htm, epub, xlsx, xls, ods, csv, pptx, ppt, odp, png, jpg, jpeg.
Merge. Up to 20 documents into one PDF, in any order.
Split. Extract specific pages or page ranges as separate documents.
Compress. Three levels: low, medium, high.
Search. Semantic search across the whole library, returning matching pages grouped by document.
Organize. Create folders, move, rename, copy and delete documents and folders.
Download. A direct download link for any document in your library.
Editing is the part most PDF integrations do not have. The assistant opens an editor session, reads the document structure (pages, paragraphs, positions) and applies targeted changes paragraph by paragraph. Supported edits include replacing or updating text in any paragraph, alignment, line height and indentation, adding and removing pages, moving, duplicating and reordering pages, adding, moving and removing text blocks, and bold, italic, superscript, subscript and custom font styling per text run.
Things to try
Generate a standard document "Generate a payment invoice: Acme LLC, $15,000, software development services, March 2026."
Fill a template for a specific client "Here's my proposal template and the client details: client Grand Pacific, contact Sarah Chen, projected revenue $42,000/mo, date today. Fill in the template and send me the result."
Batch generate personalized proposals "Here's my proposal template. Generate individual PDFs for each of these 10 clients: [paste table]. Name each file after the client and give me all the download links."
Prepare a deal package "Merge the contract, Exhibit A, and the invoice into one PDF, update all dates to today, compress to the smallest size, and give me a download link."
Edit contract details "In the contract, replace 'Old Corp' with 'New Corp' everywhere and update the effective date to May 1, 2026."
Translate a document "Translate this NDA from English to Spanish and give me a download link." The assistant rewrites each paragraph in the target language directly inside the PDF, preserving layout.
Split a report into sections "Split this 40-page report: pages 1-10 as the executive summary, 11-30 as the main body, 31-40 as appendices."
Organize your library "Create folders for Contracts, Invoices, and Reports. Move all files to the right folder by type."
Tools
Get files in
upload_file: add a local file to the library. Returns a single use PUT URL to stream the bytes to. Non PDF input is converted to PDF.import_file_from_url: fetch a publicly accessible http(s) file and store it as a document. Up to 100 MB.generate_pdf_document: create a new PDF from a natural language prompt, for example "NDA agreement" or "Sales invoice for Acme LLC, January 2026".
Transform
convert_pdf_document: export to any supported format. Returns a temporary download URL, or a ZIP when the output is multiple files such as PDF to PNG.merge_pdf_documents: concatenate 2 to 20 documents into one new document, in the order you pass them.split_pdf_document: one new document per page range. Ranges are 1 based:"1-4","7","6-z"for "to the last page". Up to 50 ranges per call.compress_pdf_document: compressed copy at low, medium or high. The original is untouched.copy_pdf_document: duplicate a document under a new name.
Read and search
find_in_documents: semantic search across the whole library or a chosen subset. Returns matching pages grouped by document.list_document_tree: full recursive JSON tree of folders and documents.download_pdf_document: pre authenticated download URL, valid 10 minutes, single use.
Organize
create_folder: new folder, at root or nested.move_resource: reparent a document or folder.rename_resource: rename a document or folder in place.delete_resource: move to trash. Folder deletes cascade. Recoverable from the pdf.net web UI.
Account and handoff
get_profile: the authenticated user profile.open_in_web_editor: deep link into the pdf.net editor for tasks the connector does not cover yet.
Limits
- 100 MB per document for import, merge and split
- 2 to 20 documents per merge, up to 50 ranges per split
- Generation prompt up to 2500 characters, search query up to 500
- Download, upload and conversion URLs expire after 10 minutes and are single use
Not available through the connector yet
E-signing and signature requests, annotations, form filling, and password protection. Ask for one of these and the assistant returns a direct link to open that document in the pdf.net editor, where you can finish the task.
Auth and security
Every tool is strictly scoped to your authenticated account. There is no way to reach another user's documents, and the connector cannot touch your email, calendar or any other connected app. All data is transmitted over TLS 1.2+ and documents are stored encrypted at rest with AES-256. We do not sell your documents or their contents, and we do not use them to train AI models. Disconnect at any time from your assistant's settings.
Links
- Product: https://pdf.net
- MCP landing page: https://pdf.net/mcp-server
- Setup guide: https://help.pdf.net/article/29-connect-pdf-net-to-your-chargpt-or-claude-mcp
- App: https://app.pdf.net
Web 聊天与 UI 分类下的更多 MCP 客户端

OpenSEO
every-appSEO research tools for AI agents: keyword research and metrics, SERP and local SERP results, domain and backlink analysis, rank tracking, and Google Search Console performance.
OPENWEBUI
open-webuiOpen WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform designed to operate entirely offline. It supports various LLM runners like Ollama and OpenAI-compatible APIs, with built-in inference engine for RAG, making it a powerful AI deployment solution.
Sample Nova Sonic Mcp
aws-samples
PoYo.ai
poyo.aiPoYo.ai gives developers one API key for 500+ AI models, including GPT Image 2, Seedance 2, Sora 2, Gemini 3, and Claude. Credit pricing, polling/webhook workflow, failed tasks not charged.
SPECMEM
jonhardwick-specPersistent semantic memory for Claude Code, chats AND code. 74+ MCP tools, local pgvector embeddings, codebase indexing with 14 language extractors, multi-agent team coordination. Self-hosted, zero cloud dependency.
评论