Ppdfia
🤖 Built for AI agents

Free PDF API & MCP server

Give your agent PDF superpowers: merge, split, rotate, watermark, number, flatten, and edit metadata over a plain REST API or the Model Context Protocol. No API key, no signup.

MCP server

Streamable-HTTP endpoint at https://pdfia.com/mcp. Add it to Claude Code, Claude Desktop, Cursor, or any MCP client:

{
  "mcpServers": {
    "pdfia": {
      "type": "http",
      "url": "https://pdfia.com/mcp"
    }
  }
}

Exposed tools: merge_pdfs, split_pdf, extract_pages,remove_pages, rotate_pdf, watermark_pdf,add_page_numbers, get_pdf_metadata, set_pdf_metadata,flatten_pdf. PDFs are passed and returned as base64 strings.

REST API

JSON in, JSON out; files travel as base64. The index at/api is machine-readable, so agents can discover every operation without this page.

# Merge two PDFs (files as base64)
curl -X POST https://pdfia.com/api/v1/merge \
  -H "Content-Type: application/json" \
  -d '{"files": ["<base64 of a.pdf>", "<base64 of b.pdf>"]}'
# → {"file": "<base64 of merged.pdf>"}

# Watermark every page
curl -X POST https://pdfia.com/api/v1/watermark \
  -H "Content-Type: application/json" \
  -d '{"file": "<base64>", "text": "CONFIDENTIAL", "opacity": 0.15}'

# Read metadata
curl -X POST https://pdfia.com/api/v1/metadata \
  -H "Content-Type: application/json" \
  -d '{"file": "<base64>"}'
POST /api/v1/…Does
mergeCombine 2+ PDFs into one, in order
splitSplit by comma-separated ranges (“1-3, 5”), or every page
extractPull selected pages into a new PDF
removeDelete selected pages
rotateRotate all or selected pages by 90/180/270°
watermarkDiagonal text stamp with opacity, size, color
page-numbersStamp page numbers (position, start, format)
metadataRead document properties, or rewrite/strip them
flattenBake form-field answers into page content

Limits & privacy