PDF Files

Upload PDF documents to train your chatbot on manuals, guides, and other documents.

Uploading PDFs

1. Go to Sources > Add Source
2. Select File
3. Click to upload or drag and drop your PDF
4. Wait for processing to complete

Supported Formats

File Requirements

• Maximum file size: 10 MB per file
• Maximum pages: 200 pages per PDF
• Text-based PDFs: Must contain selectable text
• Language: English (other languages in beta)

How PDFs Are Processed

1. Text extraction - All text is extracted from the PDF
2. Chunking - Content is split into meaningful sections
3. Indexing - Sections are indexed for search
4. Ready - Content available for chat responses

Best Practices

Document Quality

For best results:

• Use PDFs with clear formatting
• Ensure headings are properly formatted
• Avoid PDFs that are mostly images
• Use text-based charts when possible

Document Organization

• Use descriptive file names
• One topic per PDF when possible
• Keep PDFs under 50 pages for faster processing
• Remove unnecessary pages (cover pages, blank pages)

Content Tips

• Include context and definitions
• Use complete sentences
• Avoid abbreviations without explanations
• Ensure technical terms are defined

Updating Documents

PDFs cannot be refreshed - to update:

1. Delete the old source
2. Upload the new version
3. Wait for processing

Troubleshooting

Upload Failed

Possible causes:

• File too large (>10 MB)
• Corrupted PDF
• Unsupported PDF format
• Network timeout

Solutions:

• Compress the PDF to reduce size
• Re-export the PDF from the source application
• Split large PDFs into smaller files

No Content Extracted

Possible causes:

• Scanned PDF (image-based)
• Password-protected PDF
• PDF contains only images

Solutions:

• Use OCR software to create a text-based PDF
• Remove password protection before uploading
• Export with text layer from source application

Poor Quality Responses

Possible causes:

• Content poorly formatted
• Missing context
• Technical jargon undefined

Solutions:

• Reformat the document for clarity
• Add introductory context
• Include a glossary section