Academic research search across five databases with PDF vector & multiple exports

This workflow contains community nodes that are only compatible with the self-hosted version of n8n.

Description:

Unified Academic Search Across Major Research Databases

This powerful workflow enables researchers to search multiple academic databases simultaneously, automatically deduplicate results, and export formatted bibliographies. By leveraging PDF Vector's multi-database search capabilities, researchers can save hours of manual searching and ensure comprehensive literature coverage across PubMed, ArXiv, Google Scholar, Semantic Scholar, and ERIC databases.

Target Audience & Problem Solved

This template is designed for:

• Graduate students conducting systematic literature reviews
• Researchers ensuring comprehensive coverage of their field
• Librarians helping patrons with complex searches
• Academic teams building shared bibliographies

It solves the critical problem of fragmented academic search by providing a single interface to query all major databases, eliminating duplicate results, and standardizing output formats.

Prerequisites

• n8n instance with PDF Vector node installed
• PDF Vector API credentials with search permissions
• Basic understanding of academic search syntax
• Optional: PostgreSQL for search history logging
• Minimum 50 API credits for comprehensive searches

Step-by-Step Setup Instructions

1. Configure PDF Vector Credentials

   • Go to n8n Credentials section
   • Create new PDF Vector credentials
   • Enter your API key from pdfvector.io
   • Test the connection to verify setup

2. Import the Workflow Template

   • Copy the template JSON code
   • In n8n, click "Import Workflow"
   • Paste the JSON and save
   • Review all nodes for any configuration needs

3. Customize Search Parameters

   • Open the "Set Search Parameters" node
   • Modify the default search query for your field
   • Adjust the year range (default: 2020-present)
   • Set results per source limit (default: 25)

4. Configure Export Options

   • Choose your preferred export formats (BibTeX, CSV, JSON)
   • Set the output directory for files
   • Configure file naming conventions
   • Enable/disable specific export types

5. Test Your Configuration

   • Run the workflow with a sample query
   • Check that all databases return results
   • Verify deduplication is working correctly
   • Confirm export files are created properly

Implementation Details

The workflow implements a sophisticated search pipeline:

1. Parallel Database Queries: Searches all configured databases simultaneously for efficiency
2. Smart Deduplication: Uses DOI matching and fuzzy title comparison to remove duplicates
3. Relevance Scoring: Combines citation count, title relevance, and recency for ranking
4. Format Generation: Creates properly formatted citations in multiple styles
5. Batch Processing: Handles large result sets without memory issues

Customization Guide

Adding Custom Databases:

// In the PDF Vector search node, add to providers array:
"providers": ["pubmed", "semantic_scholar", "arxiv", "google_scholar", "eric", "your_custom_db"]

Modifying Relevance Algorithm: Edit the "Rank by Relevance" node to adjust scoring weights:

// Adjust these weights for your needs:
const titleWeight = 10;    // Title match importance
const citationWeight = 5;  // Citation count importance
const recencyWeight = 10;  // Recent publication bonus
const fulltextWeight = 15; // Full-text availability bonus

Custom Export Formats: Add new format generators in the workflow:

// Example: Add APA format export
const apaFormat = papers.map(p => {
  const authors = p.authors.slice(0, 3).join(', ');
  return `${authors} (${p.year}). ${p.title}. ${p.journal || 'Preprint'}.`;
});

Advanced Filtering: Implement additional filters:

• Journal impact factor thresholds
• Open access only options
• Language restrictions
• Methodology filters for systematic reviews

Search Features:

• Query multiple databases in parallel
• Advanced filtering and deduplication
• Citation format export (BibTeX, RIS, etc.)
• Relevance ranking across sources
• Full-text availability checking

Workflow Process:

1. Input: Search query and parameters
2. Parallel Search: Query all databases
3. Merge & Deduplicate: Combine results
4. Rank: Sort by relevance/citations
5. Enrich: Add full-text links
6. Export: Multiple format options