Upload your data. Ask it anything.

Every spreadsheet starts the same way. You get the file. You open it. You stare at it for a minute trying to remember whether you want a pivot table or a SUMIFS. You build something, it doesn't quite answer the question, you rebuild it, and 20 minutes later you have an answer you could have gotten in 30 seconds if you could just ask the file directly.

Data Chat is that. You upload the file, you ask the question, it answers.

What it handles

Anything you'd normally ask a spreadsheet with a pivot table, a formula, or a chart. Totals by category. Trends over time. Top N lists. Segment comparisons. Outliers. Correlations between columns. Forecasts and moving averages for the more ambitious stuff.

It also handles the follow-up questions. Most tools break the second you ask a clarifying question because they lose context. Data Chat remembers what you uploaded, what you already asked, and what it already told you, so “what about just the West region?” actually works.

File formats

CSV, Excel (.xlsx, .xls), JSON, Parquet. Multiple files at once if you want to join or compare across them. Files stay yours — nothing gets sent anywhere it doesn't need to go.

Where it beats ChatGPT

Two places, mainly. First, it's built around the data, not around being a general chatbot. The UI shows the table, the chart, the SQL query it generated, all side-by-side, so you can actually verify the answer. Second, it runs the analysis in your browser using DuckDB WASM — your file doesn't leave your machine unless you explicitly share it.

Both of those matter if you're working with anything confidential.

Where it beats a traditional BI tool

Traditional BI tools need a data model, a dashboard build, and usually a data engineer. Data Chat needs a file. The tradeoff: it's not a governed dashboard for 200 people, it's a sandbox for one person exploring a question. Different tools, different jobs.

Real example

A marketing ops person uploads last quarter's campaign export from Salesforce. They ask: “Which campaigns had the best pipeline velocity?” The tool returns a ranked list, explains how it defined velocity (days from lead creation to opportunity conversion), and offers to chart it. Follow-up: “Break that out by source.” Done. Follow-up: “Now show me just the ones where we spent more than $5k.” Done.

Three minutes, no SQL, no pivot table, no formula.

Who this is for

Analysts who don't want to write SQL for every ad-hoc question. Marketers who pull exports and need to make sense of them fast. Ops people who live in spreadsheets but aren't power users. Anyone who's ever wished they could just talk to their data.

What this isn't

It's not a replacement for a data warehouse. It's not a BI dashboard. It's not going to handle 50M-row datasets (keep those in a warehouse). It's for the “I just got this file and I need to understand it” use case, which turns out to be 80% of the analysis work most people do.