Pricing
A seat is $20. That is the least interesting thing on this page.
A seat is an email address that can put a question to our system and get an answer back it can stand behind. Email it, or let the AI agent you already use ask on your behalf. Either way the statistics are real, the method is named, and the working comes with it.
What a seat is
An email address, and several ways to use it.
Not a login you have to remember and not a dashboard you have to learn. A seat is an address our system answers, and you reach it however suits the question, so the work happens where you already are.
In your account
Upload the file, say what you want to know
The way most people start. Bring a spreadsheet or point at a URL, describe the objective in a sentence, and watch the answer build. No method to choose.
How it works →From your AI chat
The agent you already use can ask for you
Claude, ChatGPT, Cursor or your own client connects over MCP. Your agent brings the data and the objective, we do the statistics, and it hands the result back in the conversation.
Connect your agent →From your live data
Point it at the source instead of a file
Connect Google Analytics, Search Console and other sources once, and ask questions of live data without exporting anything. The same seat, no upload step.
See the connectors →Or before you sign up
Try it on your own data with no account
A set of the library's analyses run in the browser with no login at all. Bring a CSV, get the full report, decide afterwards whether a seat is worth it.
Open the free tools →One more, and it needs no software at all: email your data and your question to [email protected] like you would a colleague, and the answer comes back in the thread.
What every seat buys
Four things, and they do not come in tiers.
These are the whole promise, as true of the first question you ask as the hundredth. No plan holds one of them back to sell you the next one.
Simple
You bring the objective, not the method
Say what you want to know in your own words. Choosing the test, the model and the assumptions is our job, and getting that choice wrong is the most expensive mistake in applied statistics.
How it works →Repeatable
The same question next quarter gives the same analysis
Real computation in R, not a language model's guess, so the numbers do not drift between months. This quarter stays comparable to last, which is the whole reason to run it again.
What reproducible means here →Defendable
The method is named and the code comes with it
Every answer arrives with what it did and the source that did it. Someone who was not in the room can check the claim instead of taking it, which is the difference between a figure and evidence.
Why we write R →Searchable
It joins a library you can ask questions across
Every answer stays findable, by you or by the agent in your chat, and cites the analyses it came from. The work compounds instead of scrolling away.
Browse the library →How far we go
Not every question needs the same machinery.
So we do not charge you as though it did. Most questions settle long before the expensive part, and you never pick the rung. The question does.
Straight answer
Some questions do not need evidence
The average of a column. A count. Something you could check by eye and nobody would ask you to defend. Running a statistical pipeline at it would be theatre, so we just answer.
Estimate
A rough answer before you commit to anything
We read a sample and come back with roughly what is there and the shape the full analysis would take. Free, about half a minute, and often where the question ends, because finding out the effect is not there is a real result.
From the library
Someone has already answered this shape of question
A prebuilt, independently verified analysis that fits your data and your objective. About two minutes, names its method, hands you the R source, and is yours to run again whenever the question comes round.
Built for you
When nothing in the library covers it
We work out the method, write the code and verify it against your question. What comes back is a durable analysis you own, so the twentieth time you ask, the work is already done.
Climbing a rung is a judgement about the question, never an upsell. If the library already answers it, that is what runs, and if it tries and cannot, we build you one rather than handing back a failure. The useful question was never how much you want to spend. It is what the answer has to survive.
What the library covers
Twelve questions, not a list of tests.
The library is organised by what you are trying to find out, not by what the method is called. You do not need to know whether you want a t-test or a mixed model, you need to know which of these you are asking.
- Is the difference real?Group differences
- Which way is this heading?Trends over time
- Can I trust this data?Data quality
- What's actually driving this?Outcome drivers
- Who are my natural groups?Segmentation
- What is my survey telling me?Survey & ratings
- Did it actually cause it?Cause & effect
- Can I trust these predictions?Model quality
- Where am I losing people?Growth & funnels
- How much data do I need?Study design
- What are the odds?Bayesian methods
- What does all the evidence say?Evidence synthesis
Each of these is a family of prebuilt, verified analyses, and your objective is matched against them before anything is built from scratch. Browse the library →
Included everywhere
The same platform on every plan.
Nothing below is withheld to make a more expensive plan look better, and the demo has what a seat has. What changes between plans is how many people are asking.
- Every analysis in the library, on the demo as well as a seat.
- A closer reading on the questions that deserve one.
- The R source and the method on every answer, so it can be checked.
- Your searchable library. Ask across everything you have ever run.
- Live connectors and standing schedules. Point at the source, skip the export.
- Failed builds are never billed.
And the price
A seat, by the month.
Business is the same seat, twenty of them, at the same rate. That is the entire pricing model; there is nothing else to read.
Academic
$20/mo
One seat
One person, and everything on this page.
- ✓Every analysis in the library
- ✓A closer reading when it matters
- ✓R source and method on each answer
- ✓Live connectors and schedules
Business
$400/mo
Twenty seats, at the same $20
A seat each, assigned by email, for a team that asks together.
- ✓Everything in Academic, per seat
- ✓A setup call: we run your first analysis with you
- ✓Priority queue
- ✓Add or drop seats any month
Enterprise
By contract
Your environment
The platform under your brand, on your cloud or servers you own.
- ✓White-labelled deployment
- ✓Pipelines customised to your recurring questions
- ✓On premise, or your cloud
- ✓Unlimited usage
Each seat carries a month of usage. If a month runs dry before the questions do, a boost buys that seat's month again, at the same $20. No volume ladder and no discount for buying more at once, because a boost is the seat again rather than a bulk purchase.
Questions
The ones people actually ask.
What's a credit?
Credits are our unit of pricing. Running an analysis spends them, and so does the review that writes it up. What each costs moves with the model that reads your results and with what the work actually takes, so we do not quote a fixed rate. You are never charged for a run that fails, and we check you can afford one before it starts.
What does it cost to run an analysis from the library?
The run, plus the review that turns the numbers into a report you can hand to someone. The library is our own prebuilt, independently verified analyses, so there is no build and no wait. The exact cost depends on the model and on the work involved. A run that fails is never charged.
What does a closer reading change?
The same analysis, read more closely. Nothing about the statistics changes: the method, the code and the numbers are identical. What changes is how much the write-up notices, and it spends more of your month. Pick it per question, not per plan.
What do I get when I sign up?
A 14-day demo: every tool in the library, no card. Prices on the demo are the same as on any plan. When the demo ends you keep everything you built, and you pick up a seat.
What's a boost?
When a seat's credits run out before the month does, a boost buys that allowance again at the same designation's rate. It is the seat again, not a bulk purchase: there is no volume ladder and no discount for buying more at once. A plan is the rhythm; a boost is the exception.
Do credits expire?
Plan credits refill each billing cycle and do not roll over. Demo credits last the fourteen days.
How long are my reports kept?
Reports are kept for at least a year, and any report you cite - generate a citation, share it, pin it, or reference it in another analysis - is retained permanently regardless of plan. Cited reports are part of the reproducibility guarantee: we never delete research you have committed to.
What are seats?
A seat is one person, with a month of usage on it. Academic is one seat at $20. Business is twenty seats at $400, which is the same $20 a seat, plus a setup call, a priority queue and live connectors. Past twenty it is $20 a seat. Seats are mainly about how many people are asking.
When does a seat make sense?
When the question comes back. A seat refills every cycle, which is the difference between asking once and having a standing analyst. If you run short in a given month, a boost covers it without changing the plan.
What's Enterprise?
Enterprise is the platform becoming yours, not a bigger Business plan. Two ways in. Plug and play: we deploy the platform white-labeled inside your environment, on your cloud or servers you own, with pipelines customized to your business. Or build your own: you are building your own agentic pipelines and we consult and co-build alongside your engineers. Unlimited usage, priced by contract. Start at /enterprise or book a call.
Can I share an analysis I own?
Yes, via 7-day links. Recipients run it on their own data and pay their own credits; you are not charged for their usage.
Not sure yet
Send the question first. See what comes back, then decide.
Cymple
Data Scientist
Send me your data and question, I’ll send you the analytics.