PaperDog API Documentation - REST & MCP Endpoints

Overview

PaperDog provides both REST API and Model Context Protocol (MCP) endpoints for accessing AI research papers from arXiv and HuggingFace.

Base URL

https://paperdog.org

Authentication

No API key required. Rate limited by IP address.

REST API Endpoints

Get Papers

GET

Endpoint: /api/papers

Description: Get list of papers with pagination and filtering

Query Parameters:

page (number) - Page number (default: 1)
limit (number) - Results per page (default: 20, max: 100)
category (string) - Filter by category
search (string) - Search query

curl "https://paperdog.org/api/papers?limit=10&category=machine_learning"

Search Papers

GET

Endpoint: /api/search

Description: Search papers with advanced filters

Query Parameters:

q (string, required) - Search query
category (string) - Filter by category
limit (number) - Max results (default: 20)
min_score (number) - Minimum relevance score (1-10)

curl "https://paperdog.org/api/search?q=transformer&limit=5&min_score=7"

Get Paper Details

GET

Endpoint: /api/papers/:id

Description: Get detailed information about a specific paper

curl "https://paperdog.org/api/papers/12345"

Get Categories

GET

Endpoint: /api/categories

Description: Get all available research categories

curl "https://paperdog.org/api/categories"

Archive Search

GET

Endpoint: /api/archive/search

Description: Search historical paper archives

Query Parameters:

query (string) - Search query
start_date (string) - Start date (YYYY-MM-DD)
end_date (string) - End date (YYYY-MM-DD)
category (string) - Filter by category
limit (number) - Max results

Model Context Protocol (MCP)

PaperDog implements the Model Context Protocol for AI agent integration.

MCP Endpoint

Endpoint: POST /mcp

Protocol: JSON-RPC 2.0

Content-Type: application/json

Available MCP Tools

paperdog_search_papers

Description: Search across arXiv and HuggingFace papers

Parameters:

{
  "query": "string (required)",
  "category": "string (optional)",
  "limit": "number (optional, default: 20, max: 100)",
  "min_score": "number (optional, range: 1-10)"
}
                            

Example Request:

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "paperdog_search_papers",
    "arguments": {
      "query": "computer vision",
      "category": "computer_vision",
      "limit": 10
    }
  }
}
                            

paperdog_get_daily_papers

Description: Get curated daily papers with AI analysis

Parameters:

{
  "date": "string (optional, format: YYYY-MM-DD)",
  "category": "string (optional)"
}
                            

paperdog_get_paper_details

Description: Get detailed paper information including analysis

Parameters:

{
  "paper_id": "string (required)",
  "date": "string (optional, format: YYYY-MM-DD)",
  "include_analysis": "boolean (optional, default: true)"
}
                            

Code Examples

JavaScript (REST API)

// Search for papers
const response = await fetch('https://paperdog.org/api/search?q=transformer&limit=10');
const data = await response.json();

console.log('Found papers:', data.papers.length);
data.papers.forEach(paper => {
    console.log(paper.title, paper.authors);
});
                            

Python (requests)

import requests

# Search papers
response = requests.get('https://paperdog.org/api/search', params={
    'q': 'machine learning',
    'limit': 10,
    'min_score': 7
})

data = response.json()
for paper in data['papers']:
    print(f"Title: {paper['title']}")
    print(f"Authors: {', '.join(paper['authors'])}")
    print("---")
                            

MCP Client Example

// MCP request to search papers
const mcpRequest = {
    jsonrpc: "2.0",
    id: 1,
    method: "tools/call",
    params: {
        name: "paperdog_search_papers",
        arguments: {
            query: "generative AI",
            limit: 5,
            min_score: 8
        }
    }
};

const response = await fetch('https://paperdog.org/mcp', {
    method: 'POST',
    headers: {
        'Content-Type': 'application/json'
    },
    body: JSON.stringify(mcpRequest)
});

const result = await response.json();
console.log('Search results:', result.result.content[0].text);
                            

Rate Limits

PaperDog implements IP-based rate limiting to ensure fair usage:

REST API: 1000 requests per hour per IP
MCP API: 100 requests per hour per IP
Daily Limit: 10,000 requests per day per IP

Rate limit headers are included in responses:

X-RateLimit-Limit: 100
X-RateLimit-Remaining: 95
X-RateLimit-Reset: 1640995200
                            