Help

Everything you need, in one place.

Install CLAIR, bring in a spreadsheet, ask it a question and share what you find. If something goes wrong, the answer is probably at the bottom.

Getting started

Install it, open it, let it set itself up.

You need a Windows 10 or 11 PC. Setup takes a while, because the AI model downloads once, and then it runs on your own computer.

Install on Windows

Get the installer from the download page. It is about 384 MB.
Run it. It installs for your own user account, so you do not need administrator rights.
If Windows shows a blue "Windows protected your PC" screen, choose More info, then Run anyway. The installer is not code signed yet, which is why the warning appears.

Install on a Mac

The Mac version is expected in late October 2026, and it is not available to download yet. It will need an Apple Silicon Mac (any M-series chip) running macOS 15 or later. Intel Macs are not supported. To check which chip you have, choose the Apple menu, then About This Mac. Until then, CLAIR runs on Windows 10 and 11.

The first time you open it

It checks your computer. CLAIR looks at your memory and graphics and picks an AI model that fits.
The model downloads once. Depending on your computer it is about 3 to 17 GB, so it can take a while. How long depends on your connection and your hardware. This needs an internet connection.
Some features download a little more. Predict and similar features fetch an extra component pack the first time you use them. It is a further download of several hundred megabytes.
Try a demo meanwhile. Demo datasets are ready to explore while the download finishes.
You sign in with your email. CLAIR sends a 6-digit code to your address. There is no password. It is free during early access. The Terms of Use allow one device at a time and up to three devices on an account in total.

Working offline, and the licence check

Once CLAIR is set up, analysis runs without an internet connection. The free early-access licence lasts 14 days and renews automatically over the internet, so CLAIR needs to go online at least every 14 days.

If it cannot renew, CLAIR switches to read-only mode. You can still open your projects, view your data and results, and export them, but you cannot run new analysis. Go back online and it renews. If a computer is never online, a longer offline licence is available on request: write to [email protected].

By default CLAIR contacts a short, published list of places for things like licence checks, updates and model downloads. See exactly which.

Importing

Drop a file in. CLAIR tells you what it found.

Drag a spreadsheet onto the window, or use the file picker. CLAIR reads it on this computer.

These are the kinds of file it reads:

CSVExcelOpenDocumentParquetJSONSPSSStataSASSQLiteR data

Excel covers .xlsx, .xls, .xlsm and .xlsb workbooks. If a workbook has several sheets, CLAIR shows each one with its measured rows and columns so you can choose the ones worth importing. Each sheet becomes its own dataset.

What CLAIR checks on the way in

Some problems quietly produce wrong numbers. CLAIR looks for them and tells you, rather than silently fixing or ignoring them. A file with notes still imports.

A subtotal or grand total row sitting among your data gets counted as its own record. CLAIR looks for rows that read like a total and tells you where they are, without removing them for you.
A row where every cell is blank carries no information, wherever it sits in your file. CLAIR removes those automatically and tells you how many it took out.
A row carrying more values than its header can quietly lose one. CLAIR warns you and keeps those columns as text, so a damaged number never passes as a clean one.
A company name or notice printed above the real header used to be read as if it were the header itself. CLAIR looks for that kind of banner and skips past it to the real one.

There are more of these in the tips below.

Asking

Ask the way you would ask a colleague.

Open a dataset, go to Chat, and type your question in plain English.

Asking a good question

Name the thing you want and the group you care about, using words from your own column headings. "What is the average final grade for students who attended office hours?" works better than "How are grades?"
Ask one thing at a time. If you want three comparisons, ask three questions.
If CLAIR cannot match a word you used to anything in your data, it says so and names the word. It does not guess.

What "Verified" means

Verified means the number came from a calculation that actually ran on your data and passed CLAIR's automatic checks. That does not prove the question was understood the way you meant it, so the method and the query still deserve a look. Open the query, read it, and check that it matches your question.

An answer that did not come from a calculation like that is left unmarked, and a written summary or explanation is wording from the AI model, so read it with that in mind.

Seeing where an answer came from

What you can open depends on the kind of result:

Questions about your data: the rows and the calculation.
Statistical tests: the test, its inputs and its assumption checks.
Predictions: which columns the model used and how accurate it was.
Written summaries: they are written by the AI model, and they say so.

If CLAIR noticed something that affects how far to trust a number, such as a small group or a row that looks like a total, it adds a warning to the answer. That warning goes with the answer when you export it.

Exporting

A file you can hand to someone else.

Use the export button on the page you are working in.

Full Dataset Report: PDF or PowerPoint
A chat session: HTML, PDF, PowerPoint or CSV
A dashboard: PDF
A chart or figure: PNG or SVG
A codebook: HTML or Markdown

The Full Dataset Report has an export button on Chat, Dashboards and Predict.

When an answer lists rows, the HTML and CSV exports carry every row. PDF and PowerPoint print only the first rows and say how many were left out.

Tips

What to check before you trust a result.

CLAIR shows short tips like these when it starts. Each names a way a number can quietly go wrong and what CLAIR does about it.

Importing files11

Start by dragging a spreadsheet onto the window. CLAIR reads it here, on this computer.
Your spreadsheets stay on this machine. CLAIR reads them here and never sends their contents anywhere.
A cover sheet or a notes tab in a workbook is not data. CLAIR shows the measured rows and columns for every sheet so you can tell which ones are worth importing.
Each sheet you import from a workbook becomes its own dataset in CLAIR, all kept together under the same file so you can find them later.
A subtotal or grand total row sitting among your data gets counted as its own record. CLAIR looks for rows that read like a total and tells you where they are, without removing them for you.
A row where every cell is blank carries no information, wherever it sits in your file. CLAIR removes those automatically and tells you how many it took out.
A row carrying more values than its header can quietly lose one. CLAIR warns you and keeps those columns as text, so a damaged number never passes as a clean one.
A row with fewer values than your header expects simply gets its last columns left blank, and CLAIR tells you which rows came up short so you can check the source file.
A file that separates its values with a semicolon or a vertical line instead of a comma used to open as one giant column. CLAIR now works out which symbol your file actually uses before reading it.
A company name or notice printed above the real header used to be read as if it were the header itself. CLAIR looks for that kind of banner and skips past it to the real one.
When a header spans two rows, like a category name sitting above several column names, CLAIR merges them into one clear column name instead of reading the category as a stray row of data.

Chat6

Ask a question in plain English on the Chat page. When the answer came from a query CLAIR wrote and ran, you can open the query and the rows behind it.
Chat marks an answer verified only when it came from a query actually run against your table. Anything else is left unmarked.
If one row of your file is one measurement rather than one person, an average across rows is not an average across people. CLAIR says so in the answer when it spots that.
When a survey question let people pick more than one answer, counting the cells counts combinations rather than choices. CLAIR flags the answers where that matters.
A year spread across four columns is one thing stored as four. Totals across them are not totals, and CLAIR will tell you when a question runs into that.
An average over rows where some people appear more often than others leans towards those people. CLAIR looks for repeated identifiers before it answers.

Clean8

Clean lists what is wrong with a file before it changes anything: numbers stored as text, dates in several formats, the same answer spelled four ways.
Yes, YES, Y and 1 are one answer typed four ways, and they get counted as four. Clean finds those and merges them in one step.
Clean applies only the fixes it is certain about on its own. Anything that needs a judgement is shown to you first.
A scale that runs from Never to Always has an order, and merging two of its steps changes what it measures. Clean refuses that merge unless you insist.
Filling gaps with the column average makes every one of those rows look typical. If the gaps are not spread evenly across your groups, it quietly pulls the groups together.
A date that arrived as plain text will not sort, group or forecast. Convert a Column in Clean turns it back, and shows you both readings when a format like 03/04/2026 is ambiguous.
Clean can combine two datasets from the same project, either by matching a shared column between them or by stacking their rows together.
If one dataset stores its variables in rows and the other stores them in columns, merging them directly would garble the result. CLAIR spots the mismatch and turns one to match the other first.

Statistics7

Statistics builds the summary table most papers open with: counts, averages, spread and how much of each column is missing.
When a number was worked out from fewer rows than your file holds, CLAIR prints that smaller count next to it. A count that keeps shrinking is a column worth opening.
Measuring the same people twice is not the same as measuring two groups once. CLAIR checks which one you have before it compares them.
If you would have needed 300 people to see the difference you care about and you have 60, a result of "no difference" has not told you there is none.
Before you collect anything, CLAIR can tell you how many observations you would need to detect a difference of the size you expect.
Comparing a group of 400 with a group of 9 is not wrong, but the group of 9 decides how sure you can be. CLAIR reports the count behind each side.
Before comparing two groups with a t-test, CLAIR checks whether their values look like a bell curve. When they do not, it suggests a different test that does not need one.

Missing data4

Seeing what is missing in every column at once is usually a faster read of a new file than looking at the values.
Missing values do not only cost you rows. If a column is empty for a fifth of your file, every number built on it was built on the other four fifths.
Missing at random costs you certainty. Missing for a reason costs you the answer. CLAIR checks whether the gaps in one column line up with the gaps in another.
If the people who skipped a question are not like the people who answered it, the answers you have are not the whole group.

Exports & documentation4

CLAIR can write a codebook for a dataset: every column, what it holds, and either the range it covers or the values it takes.
A column description CLAIR wrote and nobody has checked is marked unreviewed in the codebook, so a reader can tell which is which.
A warning Chat adds to an answer, like a small group or dropped rows, is not left behind when you export it. The PDF, PowerPoint, HTML and JSON versions all carry the same warning.
You do not have to go through Chat to get the Full Dataset Report. The same export button is on Dashboards and Predict too.

Troubleshooting

When something does not go to plan.

Find what you are seeing and try the steps. If nothing here matches, write to us and we will help.

Windows says "Windows protected your PC"

This is Windows being careful about an installer it has not seen before. CLAIR is not code signed yet, so the warning is expected.

On Windows

Choose More info, then Run anyway. The installer carries on and installs for your own account.

I have an Intel Mac

CLAIR cannot run on an Intel Mac, whatever version of macOS it has. The Mac version, expected in late October 2026, will need an Apple chip (M1 or later) and macOS 15 or later.

To check, choose the Apple menu, then About This Mac. An Apple Silicon Mac lists a chip such as Apple M2. A Windows 10 or 11 PC works today. If you are unsure, write to us and say which computer you have.

The model download is slow or looks stuck

The first-run download is about 3 to 17 GB, depending on the model your computer needs, so a slow connection can take a long time. A firewall, VPN or antivirus on a work or university computer can also block or slow it.

Check that you are online. On a work or university computer, ask your IT team to allow the published destinations first, then open CLAIR again. A partial download does not damage anything you have already imported.

The sign-in code did not arrive

Check your spam or junk folder first, and make sure you typed the same email address your access is registered to.

Codes expire, so use the newest one. If an old code does not work, request a new code and use that. You need an internet connection to sign in.

CLAIR is read-only and will not run new analysis

The free early-access licence lasts 14 days and renews automatically over the internet. If CLAIR could not reach the licence server in that time, for example because you were offline, because Offline mode is on, or because a firewall blocked it, it switches to read-only mode. You can still open, view and export your work.

Connect to the internet and leave CLAIR open so it can renew. On a work or university computer, ask your IT team to allow the published destinations. If a computer is never online, ask for a longer offline licence at [email protected].

Answers are slow

A computer with 16 GB of memory or more is noticeably faster, and it is the comfortable baseline. Less still works, with a smaller model chosen for you.

Close other heavy apps, especially browsers with many tabs, so CLAIR has more memory to use. If one answer takes more than about twenty seconds, CLAIR offers a one-click switch to a lighter model.

CLAIR shows that its backend is offline

CLAIR runs a small background service on your computer. A crash, a forced quit or the computer sleeping can stop it while the window stays open.

If you see a Restart backend button, use it. If the message does not clear within about 45 seconds, close CLAIR completely and open it again.

How do I uninstall CLAIR?

On Windows

Open Settings, then Apps, find CLAIR and choose Uninstall. Your data folder is kept unless you choose the option to delete it.

Where is my data stored?

On your own computer, in a data folder CLAIR keeps inside your user account. Your datasets, change history, backups, exports and AI models are all there. During normal use, your files and questions are not sent to or kept on a CLAIR server.

The exact folder for your install is listed in Settings, under Privacy, in the local processing record.

Still stuck?

Write to [email protected] and tell us:

the CLAIR version (Settings, then About) and your operating system;
the steps you took;
what you expected, and what happened instead.

Before you send a screenshot or a diagnostics file, remove patient, student, client or other confidential data. Diagnostics files can contain dataset, column and file names, so look through them first. Please do not send us confidential data. If it arrives, we delete it once we notice and may ask you for a version without it.