WeTransform is built to handle the messy reality of file exchange — suppliers and clients rarely send files in the exact format you need. Here's what you can feed into WeTransform, and what happens with each type.
📄 Structured formats
These formats work out of the box. Upload your file, and WeTransform will read it directly and move on to the Match step.
📊 CSV / TXT — comma, semicolon, tab-separated files
📗 Excel (.xlsx, .xls) — standard spreadsheets with a clear header row
🔷 XML
🔶 JSON
✅ For structured files, WeTransform will automatically detect your columns and suggest a line for your headers. You confirm, then move on to Match.
🧹 Poorly structured files — Excel with merged cells or multiple tables
Got an Excel file where the data doesn't start on row 1, columns are merged, or multiple tables share the same sheet? WeTransform detects this and offers you the Autoclean option at the upload step.
Click "Try AI Autoclean" to let the AI normalize your file before the transformation begins. You can also add specific instructions to guide the extraction — or leave it blank for a default normalization.
📑 PDFs and images
PDFs and images are automatically routed through Autoclean — no button to click. WeTransform's AI reads the document and extracts the data into a clean, structured table.
You can add a prompt to guide the extraction: "Extract only the line items from the invoice" or "Ignore the header section". If the first result isn't quite right, you can adjust your prompt and run again.
🤖 Autoclean works on meaning, not formatting. Even a scanned invoice or a PDF with multi-column layout can be normalized. Once extracted, your data enters the standard transformation pipeline — you can then apply rules, fix errors, and submit as usual.
📋 Format summary
Format | Handled how |
CSV / TXT | Direct → Match |
Excel (clean) | Direct → Match |
Excel (merged cells, multiple tables) | Autoclean (optional) → Match |
XML / JSON | Direct → Match |
Autoclean (automatic) → Match | |
Images | Autoclean (automatic) → Match |
👉 What to do next
