Skip to content
MCP server

PDF Reader (PDF Agent Stack)

Reads what is inside a PDF: text, structure tree, fonts, signature fields and tags

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
CLI
Runs
Locally
Version
0.23.0
Listing updated
3 days ago

About PDF Reader (PDF Agent Stack)

PDF Reader (PDF Agent Stack) is an open-source MCP server for files and documents. It runs locally on your machine and is started by your AI client.

Most PDF servers stop at text extraction. This one inspects the file itself and reports, for each answer, what it could not read. The README says it returns what is in a PDF and never decides whether it conforms; that job belongs to pdf-verify-mcp, and spec lookups belong to pdf-spec-mcp, the other servers in the same family.

Tools run in tiers: page count, metadata, text with Y-coordinate reading order, search with context, embedded images as PNG or JPEG, remote PDF text, page rendering via PDFium-WASM, and a summary, followed by structure inspection of the object tree, catalogue, XMP declarations and signature structure. read_text reports text extractability per page, so an empty result is not mistaken for an empty page, and can resolve /ActualText replacements.

It runs on stdio from npm and is MIT-licensed. The site says the server has 20 tools, while the README says 18. The site describes the wider stack as four servers and three skills for reading, verifying and writing PDFs.

Install

Run in a terminal

claude mcp add pdf-reader-mcp -- npx -y @shuji-bonji/pdf-reader-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "pdf-reader-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@shuji-bonji/pdf-reader-mcp"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "pdf-reader-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@shuji-bonji/pdf-reader-mcp"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "pdf-reader-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@shuji-bonji/pdf-reader-mcp"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "pdf-reader-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@shuji-bonji/pdf-reader-mcp"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "pdf-reader-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@shuji-bonji/pdf-reader-mcp"
      ]
    }
  }
}

Tools this server provides (12)

Taken from the project documentation; the running server may differ.

get_page_count
Lightweight page count.
get_metadata
Full metadata extraction.
read_text
Text extraction in reading order with per-page extractability.
search_text
Full-text search with context.
read_images
Embedded images as PNG or JPEG.
read_url
Fetch a remote PDF and extract its text.
render_page
Rasterise pages to PNG or JPEG.
summarize
Quick overview report.
inspect_structure
Object tree and catalogue analysis.
inspect_catalog
Catalogue inventory: bookmarks, layers, attachments, XMP and what runs on open.
inspect_signatures
Inspect signature structure.
extract_structured_text
Text in logical order for tagged PDFs.

What it does

Text with extractability
Reports per page whether text can be extracted.
Structure inspection
Examines the object tree, catalogue, bookmarks, layers and attachments.
Image and page output
Returns embedded images or rendered pages as PNG or JPEG.
Signature structure
Inspects signature fields without verifying cryptography.
Declared gaps
Names what it could not read instead of omitting it.
Family of servers
Pairs with spec and verification servers.

Good for

  • Check whether a scan has a text layer
  • Read a tagged PDF in logical order
  • List attachments and what runs on open
  • Render a page for a vision model
  • Inspect signature fields before verification

Pricing

Open source

Prices change. Check shuji-bonji.github.io before you buy.

PDF Reader (PDF Agent Stack) FAQ

Does it verify signatures?

No; it inspects signature structure. Cryptographic verification belongs to pdf-verify-mcp.

How do I run it?

Use npx -y @shuji-bonji/pdf-reader-mcp@latest in an mcpServers entry.

Can it read a PDF from a URL?

read_url fetches a remote PDF and extracts its text only.

How many tools?

The README says 18; the site says 20.

Is PDF Reader (PDF Agent Stack) free?

Yes. PDF Reader (PDF Agent Stack) is open source, so you can use and self-host it for free.

How do I add PDF Reader (PDF Agent Stack) to Claude Code?

Run this command in your terminal: claude mcp add pdf-reader-mcp -- npx -y @shuji-bonji/pdf-reader-mcp

What are the best alternatives to PDF Reader (PDF Agent Stack)?

The closest alternatives to PDF Reader (PDF Agent Stack) are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to PDF Reader (PDF Agent Stack)

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ouroboros

    MCP server

    Spec-first layer that pins requirements and verifies results of AI coding agent work

    Runs locally
    • Open source
    • 6.4K GitHub stars
    • Developer tools