AI Data Coworker

Data

Bring data in from a file, the clipboard, or a sample set. The Coworker profiles it the moment it lands.

Drop a file here, or click to browse
Your file is read in the browser — nothing is uploaded.
.csv.tsv.txt .xlsx.xls.json

Paste a table

Copy cells from Excel or Google Sheets and paste them below.

Or try a sample dataset

Deliberately messy — good for seeing the cleaning tools work.

Clean

Every fix is recorded as a step, applied to the original data in order. Remove a step and the change is undone — nothing is destructive.

Load data first, then clean it here

Bring in a table and this becomes a full cleaning workspace: column profiles, detected problems, and fifteen operations you can apply and undo.

Model

Load several tables and link them. Columns from a linked table are looked up onto the one you are analysing, so every chart, filter and report works across the whole model.

One table so far

Add a second file — a lookup, a dimension, last year's extract — and I will scan every column pair for a key that joins them, and tell you what percentage actually matches.

Validate

Rules that say what the data is allowed to look like — then a loop that fixes what it safely can and re-checks until nothing changes.

Load data first, then set the rules

Validation checks completeness, uniqueness, ranges, allowed values, patterns, dates and referential integrity across your linked tables.

Charts

Build a chart, preview it live, then send it to the dashboard.

Load data first, then build charts here

The chart builder needs a table to work from. Bring one in and twelve chart types unlock — bar, line, scatter, pie, histogram and the rest.

Dashboard

Arrange the charts that matter. Widths are adjustable; the report and deck are built from whatever is here.

Your dashboard is empty

Build a chart yourself on the Charts tab and add it here — or let the Coworker assemble a first pass for you.

Report

Written from your data and from what you tell it the data is about. Every section is yours to keep, move, annotate or replace.

No report yet

Load data, then press Generate. The Coworker writes the findings from real statistics and embeds your dashboard charts.

Statistics

Pick the question rather than the test. The right one is chosen from what your columns actually contain, its assumptions are checked in the open, and the effect size is reported next to the p-value — because a p-value on its own has never told anyone whether something matters.

Convert

Any format in, any format out, in batch, on this device. Dates stay dates, blanks stay blank, and the quoting is done properly — which is the whole difficulty with file conversion and the reason free websites keep getting it wrong.

Extract

Notes, emails, invoices, forms, logs — anything written for a person rather than a machine. Paste it or load it, and it comes out as a table you can check, correct and then analyse like any other dataset.

Reconcile

Two files, one key, a tolerance you choose. This says exactly where they disagree — row by row, with the totals — instead of leaving you to find out by eye.

Language

Everything this app writes — the report, the deck, the PowerPoint, every chart title, subtitle and explanation — can come out in another language. The translation runs here, on this device, from a catalogue of the sentence templates the app actually produces, so the numbers, your column names and your own words are never disturbed.

Privacy

What this file does and does not do with your data, and the log to check it against.

Compare runs

Take a snapshot of this analysis, then point the same recipe at next month's export. What moved, what broke and what quietly changed shape is listed for you.

Slides

A presentation-ready summary deck. Present it full screen, or export it as a real PowerPoint file (.pptx) with editable text, as a self-contained HTML deck, or as PDF.

No deck yet

Press Generate and the Coworker turns your dashboard and findings into a slide summary you can export straight to PowerPoint.