When you need Rebel to work with a file — especially spreadsheets, CSVs, or large documents — you might instinctively drag and drop it into the conversation. That works for quick questions, but for deeper analysis, giving Rebel the file path is significantly more reliable.
What happens when you drag and drop?
When you attach a file, Rebel extracts the content client-side and embeds a text snapshot into the message. This means:
- Excel: Cell data is flattened to CSV — charts, formatting, and formulas are lost
- PowerPoint: Slide text only — speaker notes, images, and animations are stripped
- Word: Text only — images, formatting, and comments are excluded
- Large files: PDFs over 32MB are rejected; text files are capped at 5MB
The extracted content is frozen in that one message and consumes context whether Rebel needs all of it or not.
What happens when you give a file path?
When you say “Analyse the spreadsheet at Company/Q4-financials.xlsx”, Rebel reads the file from disk using its tools. This means:
- Selective reading — Rebel can read specific sheets, page ranges, or sections rather than dumping everything at once
- Full tool access — Rebel can use scripts to parse formulas, run calculations, and generate charts
- No extraction loss — the full file structure, formatting, and metadata are preserved
- Re-readable — Rebel can go back to the file on any later turn, not just the turn where you attached it
Quick comparison
| Drag & drop | File path | |
|---|---|---|
| Best for | Quick visual questions, screenshots, small text files | Spreadsheets, CSVs, large docs, repeated analysis |
| Content access | Static text snapshot | Full file with tools |
| Formulas & charts |  Lost |
 Available |
| Large files | Size limits apply | Can read in chunks |
| Multiple turns | Frozen in one message | Re-readable any time |
| File edits | Can’t write back | Can modify the file directly |
How to reference files easily
You don’t need to type the full path by hand. In the composer, use @files to search your workspace by meaning — just start typing what the file is about and Rebel will find it.
Or simply tell Rebel where to look:
“Read the CSV in my Downloads folder called march-report.csv”
When drag-and-drop IS the better choice
- Screenshots — Rebel needs the image sent directly for visual analysis
- Files not on your machine — if someone sent you something and you haven’t saved it yet
- Tiny files — a short CSV or text snippet where the overhead doesn’t matter
TL;DR
Drag-and-drop is convenient for quick, visual tasks. But for spreadsheets, CSVs, and anything you want Rebel to deeply analyse or edit — save the file and give Rebel the path. It gets full access to the content, can re-read it across turns, and can even write changes back.
Small habit change, noticeably better results. 