Skip to main content

Knowledge Base API

Read-only API endpoints for accessing PDF tagging reference data in both human-friendly Markdown and machine-friendly JSON formats

API Overview
Access comprehensive PDF tagging data derived from JSON sources including tags, attributes, properties, Matterhorn Protocol checks, and computed structure relationships.

Base URL

https://taggedpdfschool.pdfa.org

Response Formats

Markdown
JSON

Authentication

Currently unauthenticated (read-only). All endpoints are publicly accessible.

Available Endpoints

Tags
Read-only
Complete tag reference with descriptions, namespaces, and use cases

Human-friendly format for documentation

GET https://taggedpdfschool.pdfa.org/api/db/md/tagsOpen

Example Usage:

curl https://taggedpdfschool.pdfa.org/api/db/md/tags
Structure Relationships
Read-only
Computed aggregation describing parent/child tag constraints

Human-friendly format for documentation

GET https://taggedpdfschool.pdfa.org/api/db/md/structureRelationshipsOpen

Example Usage:

curl https://taggedpdfschool.pdfa.org/api/db/md/structureRelationships
Attributes
Read-only
PDF tag attributes with descriptions, types, and possible values

Human-friendly format for documentation

GET https://taggedpdfschool.pdfa.org/api/db/md/attributesOpen

Example Usage:

curl https://taggedpdfschool.pdfa.org/api/db/md/attributes
Properties
Read-only
PDF tag properties and their relationships

Human-friendly format for documentation

GET https://taggedpdfschool.pdfa.org/api/db/md/propertiesOpen

Example Usage:

curl https://taggedpdfschool.pdfa.org/api/db/md/properties
Matterhorn Protocol
Read-only
PDF/UA validation rules and accessibility compliance checks

Human-friendly format for documentation

GET https://taggedpdfschool.pdfa.org/api/db/md/matterhornOpen

Example Usage:

curl https://taggedpdfschool.pdfa.org/api/db/md/matterhorn
Important Notes

Responses are generated at request time from the in-app data layer bundled at build time

Structure relationships are a computed aggregation describing parent/child tag constraints

If you add or modify the underlying JSON assets, rebuild/redeploy the app to refresh endpoint output

All endpoints support standard HTTP methods and return appropriate status codes

Model Context Protocol (MCP) Server
Enable AI assistants to access TaggedPDF School's database through the MCP protocol

What is MCP?

The Model Context Protocol (MCP) is an open standard that enables AI assistants and development tools to access external data sources and tools. Our MCP server provides programmatic access to all PDF tagging reference data, allowing AI assistants like Cursor IDE and GitHub Copilot to query tags, attributes, properties, and Matterhorn Protocol checkpoints.

Configuration

To use the MCP server, add the following configuration to your IDE's MCP settings:

Cursor IDE

Add this configuration to your Cursor MCP settings file:

{ "mcpServers": { "tagged-pdf-school": { "url": "https://taggedpdfschool.pdfa.org/api/mcp", "transport": "http" } } }

Location: ~/.cursor/mcp.json (macOS/Linux) or %APPDATA%\Cursor\User\globalStorage\mcp.json (Windows), or via Settings → Features → Model Context Protocol

GitHub Copilot

Add this configuration to your GitHub Copilot MCP settings:

{ "mcpServers": { "tagged-pdf-school": { "url": "https://taggedpdfschool.pdfa.org/api/mcp", "transport": "http" } } }

Location: Check GitHub Copilot documentation for MCP configuration file location

Available Tools

The MCP server provides seven powerful tools for querying our database:

  • get_attribute - Get attribute details by name
  • get_tag - Get tag information and specifications
  • get_property - Get property definitions and values
  • get_checkpoint - Get Matterhorn Protocol checkpoint details
  • search_attributes - Search attributes by keyword
  • search_tags - Search tags by keyword
  • get_all_databases - Get summary of all available databases

How to Use

After configuring the MCP server, restart your IDE. You can then ask your AI assistant questions like: 'Get information about the Document tag', 'Search for attributes related to borders', or 'What does checkpoint 01-001 check for?'. The AI assistant will use the MCP tools to query our database and provide accurate, up-to-date information.

Common Use Cases

Markdown Format

  • Documentation pages
  • Human-readable reference
  • AI-readable reference
  • Educational content
  • Quick lookups

JSON Format

  • Automated tooling
  • Scripts and integrations
  • Data processing
  • API consumption