How the checks fit together

You do not have to pick checks. Press Check now and nth Labs runs every check your plan includes, and says why any of them waits. Every finding goes to one of three departments (IT, Marketing or Finance), and every check is either free on every plan or part of what a paid plan adds.

Running the checks

Open New run at the top of the console. Its card, The full check, lists every check it will run, each with its reason, and under Waiting, whatever cannot run yet. Press Run the full check and they run, one after another. The same button on the Readiness page runs the same check.

The first time, the plan starts with the Quick check: Page health, AI compatibility, Data saving and the Report in one pass, then a read of your pages for the checks that come after it. Regression check, also part of the Quick check, is skipped on the first run because there is nothing to compare with yet. The Quick check is free, and later runs compare against what it found.
After that, each press runs Regression check and Page health again, so the fixes you kept are tested and every page is measured afresh.
When your plan covers them, each press also runs Site review, First impressions (the simulated panel reading your pages) and User walkthroughs (simulated people from your market trying the product).
Anything that cannot run is named on the card with its reason, for example that it needs a paid plan.

If every run slot on the console is busy when you press, your run is queued instead of being turned away. The run screen says Queued, how many runs are ahead of yours and, once the console has timed runs like it, about how long. It starts on its own, keeps its place if you close the tab, and Stop takes it out of the queue.

To run one check on its own, open Run one check instead under the card; the ones that use AI say they need a paid plan. A department’s own page has a button for everything that department relies on, in its order. A schedule in Settings, Runs can ask the same planner (What’s due), and then runs nothing when nothing has changed.

The three departments

Every finding lands in exactly one department, chosen by the kind of problem it is rather than by the check that found it. The sidebar lists the three under Departments, each with its count of open findings and a dot for the worst of them. Each department’s page has a button, Run IT’s checks (or Marketing’s, or Finance’s), that runs everything the department relies on, with a line under it naming the checks it runs, and which of them need a paid plan.

Department	The question	What feeds it
IT	Does it actually work? Errors, accessibility, speed and security.	Page health, Data saving, Regression check, AI compatibility and Site mapping (which tries every screen and form, and maps the product), all free. On a paid plan, Site review and User simulation add the screens a stranger could not follow, trust or use.
Marketing	Does the pitch land? Claims, proof, who tries it and what stops them.	How a shared link to your product looks, from Page health, free. On a paid plan, Site review’s reading of your claims, proof and copy, and in a validation, the simulated panel’s reactions to your pages.
Finance	Will anyone pay, and what could it make? Price, who would pay and what it could make.	The way to money, from Page health, free: a price with nothing that can buy it, no way to sign up, no way to reach a person. On a paid plan, User simulation (a tester ready to buy who gets stuck) and Strategy review (whether this looks like a business, and why anyone would switch). The Lab, under Forecast, models what the product could make from assumptions it names.

Free checks, and what a plan adds

A check is free when it measures rather than judges. Page health, AI compatibility, Data saving, Regression check, Staging vs live, the Report and Site mapping need no AI to read your product, so they are free and unlimited on every plan, Free included.

A paid plan adds the work that reads your product the way a person would, which takes judgement and needs AI:

Site review reads every page like a first visit.
User simulation sends a simulated tester through the product.
Strategy review, in Finance’s checks, judges whether this looks like a business, and why anyone would switch.
Validations. A validation is one run of the whole readiness check: the panel of simulated buyers and the walks through your product. Its answers fill in the Readiness page, and the people it seats appear on the People page.

All of it runs on nth Labs’ AI and shares your plan’s monthly fair use, which pricing counts in validations. On Free, one full check each month includes the AI, with no card. Plans and billing says what each plan includes and what happens when you reach it.

If you run nth Labs yourself from the repository, these checks use an AI key you set on your own machine instead of a plan. See AI and your plan.

All checks

Under A single check you can run any of these on its own, and the Quick check and Full check too.

Check	What it does	Cost
Page health	Opens every page in a real browser at desktop and phone widths: speed, accessibility, broken links and images, errors, security basics, and buttons that do nothing when pressed.	Free
AI compatibility	Checks that an AI assistant acting for a person could use your product: buttons with names, everything reachable by keyboard, forms it can read, and no CAPTCHA wall in the way.	Free
Data saving	For each round trip declared for your product, types a marked value into a form, saves it, comes back and reads it out again, to prove the save held.	Free
Regression check	Runs every fix you kept in a real browser, and compares every page with the last look you accepted.	Free
Staging vs live	Opens the same pages at two addresses, such as staging and live, and shows the differences in pixels and words. Files no findings: you decide which side is right.	Free
Report	Rolls every run and finding into one document you can send.	Free
Site review	Reads every page like a first visit: clarity, proof, trust, the copy line by line, and a launch score.	Paid plan
User simulation	Sends a simulated tester, or several, into the product with no script, and reports where they got stuck.	Paid plan

Checks that run other checks

Quick check. Page health, AI compatibility, Data saving, Regression check and the Report in one pass, then a read of your pages. Free.
Full check. The same pass with Site review added, and one verdict at the end. Needs a paid plan, because of Site review. User simulation and Staging vs live are not part of it.
A department’s checks. Everything one department relies on, from that department’s page. The free checks in it run on any plan; the rest wait for a paid plan and say so.

What to run when

After each change you ship: press Check now. The plan checks what you already fixed and measures every page again.
After fixing one finding: open it and press Fixed? Verify it. nth Labs checks that finding alone, and closes it only if the fix held.
After a redesign or a new page: Site review, then User simulation on the flow that changed.
Before a staging build goes live: Staging vs live, with the second address under Compare against.
Before you show the product to anyone: read the Report, and on a plan with validations, the Readiness page.

From the command line

If you run nth Labs yourself from the repository, every check is a command, and so are the ones that run others:

node cli.cjs checkup     # the Quick check, free, in one pass
node cli.cjs sweep       # Full check
node cli.cjs validate    # the whole readiness check; a step it cannot run is named, with the reason
node cli.cjs watch       # Page health again whenever you save a file

Every command is listed on the command line page.