Analyze a CSV file with AI — ask questions in plain English

What this adds to an uploaded file — and what it does not

An uploaded file is a snapshot: we store the copy you send and analyse that, so when the underlying data changes you upload it again. What it adds over pasting the same file into a general-purpose chatbot is that on complex questions it shows a readable plan before running anything, on every answer the exact SQL and the rows it returned are one click away, and it saves what your fields mean against that source — so the next question already knows them instead of guessing. If you want answers that refresh without re-uploading, connect a Google Sheet or a database instead; the Google Sheets connection works on the Free plan.

CSV file: things to know

• A clean header row and consistent types give the best results; avoid blank header cells.
• File upload is a Pro feature; on Free, a Google Sheet is the quickest way to try file-style data.

Example questions to ask your CSV file data

• Total and average of the amount column by category
• Top 10 rows by value
• Count of records per month from the date column

Keeping it safe

• Your uploaded data is read for analysis only and never written back.
• On our side every query is checked before it runs and refused if it contains a write statement — DROP, DELETE, UPDATE, INSERT and fourteen others, eighteen keywords in all. It is a keyword guard, not a full SQL parser, so it can only refuse what it can spell. What actually keeps your file unchanged is that we analyse a stored copy of it; the guard is the second line, not the first.
• Connection details and your AI key are encrypted at rest.
• We train no model on your data — we do not have one; the analysis runs on your own provider key, so the terms that decide what happens to what you send are your provider's, not ours. An uploaded file is stored in your own account so the analyst can query it, and the rows an answer returns stay there until you delete them.