n8n-nodes-pdfvector

n8n community node for PDF Vector — parse PDF, Word, Excel, and image documents to structured markdown, ask AI questions about documents, extract structured data with JSON Schema, and search academic papers across multiple databases.

Installation

Follow the installation guide in the n8n community nodes documentation.

In n8n, go to Settings > Community Nodes and install:

n8n-nodes-pdfvector

Credentials

1. Sign up at app.pdfvector.com
2. Go to your PDF Vector instance settings to create a current API key
3. In n8n, create a new PDF Vector API credential and paste your API key

Migrating an older credential: Old PDFVector API keys no longer work with the current API. Create a new API key from the PDF Vector dashboard, then replace the key in your n8n credential.

Custom instance: If you have a custom PDF Vector instance, enter your instance domain or base URL in the credential's Domain or Base URL field. Otherwise, leave it as global.pdfvector.com. Pasted URLs such as https://global.pdfvector.com/api are normalized automatically.

Operations

Document

Parse, ask, and extract from generic documents (PDF, Word, Excel, CSV, Image).

| Operation | Description | |-----------|-------------| | Parse | Extract text content and page count | | Ask | Ask a question about the document and get an AI answer | | Extract | Extract structured data using a JSON Schema |

Document Parse includes an optional Include Pages toggle. Enable it to return page-separated markdown in the pages array while still receiving the full document markdown.

Identity Document

Parse, ask, and extract from identity documents (passports, driver's licenses, ID cards).

| Operation | Description | |-----------|-------------| | Parse | Extract text from identity documents (Pro/Max models only) | | Ask | Ask a question about the identity document | | Extract | Extract structured fields (name, DOB, document number, etc.) |

Invoice

Parse, ask, and extract from invoices.

| Operation | Description | |-----------|-------------| | Parse | Extract text from invoices (Pro/Max models only) | | Ask | Ask a question about the invoice | | Extract | Extract structured invoice data (number, date, items, totals) |

Bank Statement

Parse, ask, and extract from bank statements.

| Operation | Description | |-----------|-------------| | Parse | Extract text from bank statements (Pro/Max models only) | | Ask | Ask a question about the bank statement | | Extract | Extract structured bank statement data |

Academic

Search, fetch, and find citations across academic databases.

| Operation | Description | |-----------|-------------| | Search | Search for papers across 7 academic databases | | Fetch | Fetch specific papers by DOI, PubMed ID, ArXiv ID, etc. | | Find Citations | Find relevant citations for a paragraph of text |

Supported providers: Semantic Scholar, PubMed, ArXiv, Google Scholar, OpenAlex, ERIC, Europe PMC.

Model Tiers

All document operations support model selection:

| Tier | Best for | Max pages | Max size | |------|----------|-----------|----------| | Auto | Automatic selection (default) | 1000 | 500MB | | Nano | Simple text documents | 30 | 10MB | | Mini | Tables, structured content | 30 | 10MB | | Pro | Complex docs, images, handwriting | 30 | 40MB | | Max | Large docs, full capabilities + HTML | 1000 | 500MB |

Identity, Invoice, and Bank Statement parse operations only support Auto, Pro, and Max.

Document ID Tracking

All document operations (except Academic) support an optional Document ID field for usage tracking. The ID is returned in the response and saved in your usage records.

Compatibility

• n8n version: 1.0+
• Node.js version for local development and CI: 22+

Resources

• PDF Vector API Reference
• PDF Vector Dashboard
• PDF Vector Website
• Report Issues

Version History

• 2.x: Unified PDF Vector node with document, identity, invoice, bank statement, and academic operations
• 1.x: Earlier PDF Vector integration releases

License

MIT