n8n-nodes-file-metadata

n8n node for extracting metadata from files with namespace support for Qdrant filtering.

Features

• Extract metadata from PDF files
• Extract EXIF data from images
• Process ebooks (EPUB)
• Extract archive information (ZIP)
• Parse Word documents
• Read Excel spreadsheets
• Get audio metadata
• Extract video information
• Parse markdown frontmatter
• NEW: Automatic namespace generation for Qdrant vector store filtering

Namespace Support

This version automatically adds a namespace field to the metadata based on the document title:

• Extracts title or info.Title from document metadata
• Sanitizes the title to create a valid namespace (alphanumeric and underscores only)
• Limits namespace length to 96 characters for Qdrant compatibility
• Enables filtering with queries like:

{
  "must": [
    {
      "key": "metadata.namespace",
      "match": {
        "value": "Writing_521A_Creative_Writing"
      }
    }
  ]
}

Installation

npm install [email protected]

Usage

1. Add the File Metadata Extractor node to your workflow
2. Connect it to your document source
3. Configure the binary property name (default: 'data')
4. The node will output metadata including the new namespace field
5. Use the namespace for filtering in Qdrant vector stores

Example Output

{
  "title": "Writing 521A Creative Writing",
  "author": "John Doe",
  "fileType": "PDF",
  "numberOfPages": 25,
  "namespace": "Writing_521A_Creative_Writing",
  "creationDate": "2024-01-15T10:30:00.000Z"
}

License

MIT