Uploading Files

Instead of copying and pasting content, you can upload files directly to your knowledge base.

Supported File Types

How to Upload Files

1. Navigate to Train in the sidebar
2. Click the + Add Content button
3. Select the File tab in the dialog
4. Click to browse or drag and drop files
5. Files will upload and begin processing

Processing Files

After upload, files go through these stages:

1. Uploading - File is transferred to Goldilocks
2. Extracting - Text content is extracted from the file
3. Processing - Content is chunked and embedded
4. Active - Ready to use for AI responses

Processing time depends on file size:

• Small files (< 10 pages): A few seconds
• Medium files (10-50 pages): 10-30 seconds
• Large files (50+ pages): 1-2 minutes

File-Specific Notes

PDF Files

• Text-based PDFs work best
• Scanned documents (images of text) are not supported
• Tables may not extract perfectly—consider reformatting
• Headers and footers are included in extraction

Word Documents

• Formatting (bold, italic) is stripped—only text is kept
• Tables are converted to plain text
• Images are not extracted
• Use .docx format (not older .doc)

CSV Files

CSV files are treated specially:

• Each row becomes a separate document
• First row should be headers
• Use for bulk importing FAQs or structured data

Example CSV structure:

title,content
"Return Policy","Our return policy allows returns within 30 days..."
"Shipping Info","We ship to all 50 states..."

Text Files

• Simplest format—content is used as-is
• UTF-8 encoding recommended
• Line breaks are preserved

File Size Limits

Bulk Upload Tips

When uploading many files:

1. Organize first - Use clear file names
2. Check formatting - Preview text extraction in a tool first
3. Start small - Upload a few files and verify quality
4. Use CSV - For structured data, CSV is more reliable

After Upload

Once files are processed:

1. Review the extracted content for accuracy
2. Edit titles if auto-generated names aren't clear
3. Remove any documents that didn't extract well
4. Test with the Search feature

Troubleshooting

"File type not supported"

Check that your file has the correct extension (.pdf, .docx, .txt, .csv).

"File too large"

Reduce file size by:

• Splitting into multiple files
• Removing images/graphics from Word docs
• Compressing PDFs

Poor text extraction

If extracted text is garbled or missing:

• PDF may be scanned (image-based)—not supported
• Try a different export format
• Copy/paste content manually instead