Skip to content

MaraponeAI · the mobile app

iOS & Android · Q4 2026

THE WHOLE SUITE.
IN YOUR POCKET.

Takeoff, code audit, estimating, bid leveling, buyout and spec conflicts — the same five engines that run in the browser, reached from a truck, a trailer or a scaffold instead of a desk. Not a viewer. Not a companion. The tools themselves.

Monthly or yearly, cancel any time. Already on a plan? The phone is included at no extra cost, on the same login.

See what it costs ↓ I am already on a plan → Tell me when it lands
5 products 43 jurisdictions 9 currencies Cancel anytime

Every screen on this page is the real build, recorded against real drawings and real quotes. Nothing is a mockup.

The part we are not going to dress up

ONE PLAN.
BOTH SCREENS.

Your plan covers the browser and the phone. Same account, same engines, same numbers — there is no second subscription and nothing extra to buy. What is different on a handset is physical, not commercial, and it is set out plainly below.

01

A PHONE CANNOT RUN THIS.

The five programs are roughly 48,000 lines of Python and our own model, sized for a laptop with 16 GB of memory. There is no version of that which runs on a handset — and handsets are not one machine anyway. A four-year-old Android and a new iPhone would give two different answers to the same takeoff, which is worse than giving none.

02

SO THE WORK RUNS ON A SERVER.

The phone is the screen; the five engines run on our infrastructure and hand back the answer. That is a service with a monthly cost to us — compute, storage, the model, keeping 43 jurisdictions current. Charging once for something that costs every month is how software gets abandoned two years later.

03

SO IT IS PRICED NOT TO REPLACE OWNING.

One subscription covers every screen you sign in on — the phone in your pocket and the browser on the estimator’s desk — so there is nothing to weigh up here and nothing to buy twice.

Which means one honest difference

A document you audit is uploaded to be audited — on the phone and in the browser alike. What we promise is where it stops. It goes to our server and no further: the model reading it is our own, not OpenAI's or Google's, each account gets its own isolated workspace, and nothing you upload trains anything. If your contract requires that documents never leave your own infrastructure, say so before you subscribe — deployment into your own cloud tenant is an enterprise arrangement, not a plan. How the data is held →

If you have already bought

YOU DO NOT
PAY TWICE.

Your plan is not a browser plan. Sign in on the phone with the same email and the programs your plan covers are open there too — Field opens two, Pro and Firm open all five. There is no mobile tier, no second invoice and no code to redeem.

  1. 01 Start a plan. In the browser, on the pricing page. One subscription, whatever you sign in on afterwards.
  2. 02 Install MaraponeAI on your phone. Free from the App Store or Google Play. Open it and sign in.
  3. 03 That is it. The programs your plan covers are already there, with the same projects and the same rates. Nothing to enter, nothing extra to pay.

Access is tied to your account, not the handset. Change phones, reinstall, switch from iPhone to Android — sign in and it is there. A seat is a person, so the same seat covers the office desktop, a laptop and a phone without anyone counting devices.

Start on Pro — $249/mo Compare all four plans
The MaraponeAI account screen, showing the subscription tier, the city the rates are coming from, and the jurisdiction in effect — HST 13%, Ontario Building Code, 10% holdback.

What is actually in it

FIVE PRODUCTS.
NOT FIVE PREVIEWS.

No engine was rewritten for the phone. The mobile app drives the same code the desktop drives, so a bid priced on a phone and the same bid priced on the laptop are the same number — and a fix to a desktop engine reaches the app with no porting step. How that works →

Blueprint Auditor

PHOTOGRAPH THE DRAWING.
GET THE CODE REVIEW.

Point the camera at a sheet, or pull the set out of the folder it already lives in. The engine reads it, measures the rooms, extracts the quantities and checks it against the building code that applies where the job is — and marks what it found on the sheet, so you are looking at the drawing rather than a list about it.

51 code checks against the Ontario Building Code, with the counts of what was evaluated, passed, skipped and flagged for manual review.
51 checks, against the code where the job is.
Fifteen rooms read off the drawing with their use and area in square feet.
Fifteen rooms, measured and classified.
Quantities taken off the drawing — concrete slab, footing, framing lumber, wall and roof sheathing.
Quantities, ready to price.
The drawing itself on the phone, with findings pinned to the place on the sheet they were found.
Findings pinned to the sheet.

AI Estimator

A PRICE YOU CAN
DEFEND IN A MEETING.

Quantities in, a bid out — with the whole build-up visible: direct cost, escalation, general conditions, contingency, overhead, profit, insurance and the tax rate for the city you are in. Then the part estimators actually argue about: a risk register with what each risk is worth, a Monte Carlo run, and a margin ladder showing what raising your number does to your odds of winning.

A bid of CA$526,796 on a 2,400 square foot addition, with the full build-up from direct cost to tax.
The bid, and every number under it.
The scope tab, listing assemblies by CSI division with their quantities.
Every assembly, by division.
A recommended contingency of 3.0% with the five risk drivers that produced it.
Contingency, with its reasons.
Bid strategy — recommended bid, expected-value optimum, risk-adjusted floor and a margin ladder with win probabilities.
What each margin does to your odds.

Bid Leveler

THE LOW BID
IS USUALLY NOT THE LOW BID.

Photograph the quotes, forward the emails, drop the spreadsheets. Every bid gets pulled onto one scope, and what each one quietly left out gets priced back in. In the run below, the number at the bottom of the page submitted CA$1,142,000 — and landed fourth once everyone was carrying the same work.

Five mechanical bidders ranked by their levelled position, each showing what was submitted and what was added.
Ranked once the scope matches.
Forty scope gaps between the bids — controls and building automation, overtime and weekend work, the fire pump and jockey pump.
Forty gaps, named and priced.

ScopeGuard

WHAT NOBODY PRICED,
BEFORE YOU AWARD IT.

Twenty trades read against the project manual, and every requirement nobody carried flagged with what it is worth bought now versus bought as a change order after mobilisation. Winter concreting excluded by the forming contractor. The fire pump nobody mentioned. Each one cites the section it came from, so the clarification letter writes itself.

SpecChecker

WHERE THE MANUAL AND
THE DRAWINGS DISAGREE.

Fire ratings that do not match. Glazing specified one way and drawn another. Each conflict carries both citations and the reason it matters, and turns into an RFI in one tap. It also tells you what it could not check — where the drawings are silent there is nothing to disagree with, and silence looks exactly like agreement on a findings list.

Why five in one app beats five apps

ONE JOB,
CARRIED ALL THE WAY.

A drawing audited on Monday prices itself on Tuesday without anyone re-typing a quantity. The estimate that comes out of it is the one the quotes get levelled against, and the scope you bought out is checked against the same manual. The app keeps the thread and tells you what is next.

  • Measure the drawings — then hand the takeoff straight to the estimator
  • Price it against the rates for the city the job is actually in
  • Cross-check the manual against the drawings before anything is ordered
  • Level the quotes onto that scope
  • Close the buyout with what nobody priced already flagged

The five engines share one workspace per account. No integration to configure, and no export-import step in the middle.

A job on the phone showing the chain: the drawings measured, the job priced, the documents cross-checked, the quotes levelled, and the buyout still open with a button to start it.
The city list in the app — Canadian, American and European cities, each with its own currency.

Not a Toronto product with a currency switch

PRICED FOR THE CITY
YOU ARE ACTUALLY IN.

43 jurisdictions ship in the app — 12 Canadian, 12 American, 19 European, across 9 currencies. Each one carries its own tax, building code family, statutory holdback, frost depth, winter productivity loss and per-cost-type city index. Pick Vancouver and the holdback becomes the BC Builders Lien Act. Pick Copenhagen and the estimate carries 25% Danish VAT, and says so.

Verified

Tax rates, currency, measurement system, applicable code family. Published, stable, checkable.

Indicative

City cost indices, labour burden, escalation. Market data — good to price against, not law.

Calibrate

Permit fees and development charges. Set per municipality, so they ship blank — the app prices them at zero and raises a flag rather than inventing a number.

The risk register moves with you too. Of its 26 rules, ten are Ontario-bound: five are universal with a local citation, and five are artefacts of Toronto by-laws that get dropped outside Ontario — with the contingency recalculated and the removal stated on screen rather than quietly applied. The rate and code packs →

What it costs

FOUR PLANS.
CANCEL WHENEVER.

These are the same four plans that open the programs in your browser — one subscription covers both, and there is no separate app tier. Yearly is billed at ten months, so committing for a year is two months free rather than a vague saving. No contract, no minimum term, no cancellation fee, and it runs to the end of what you paid for.

Field

Measure it and price it, one seat.

$99/month

or $990 a year — two months free

  • Blueprint Auditor and the AI Estimator
  • 1 seat — phone, laptop and desktop
  • Unlimited projects and documents
  • Full exports — PDF, Excel, CSV
  • Every supported region and currency

For the estimator working alone, or a firm testing us on one live job.

Pro

All five tools

All five programs, for a team of five.

$249/month

or $2,490 a year — two months free

  • All five programs
  • 5 seats
  • Unlimited projects
  • Document intelligence and OCR
  • Monte Carlo, risk and bid strategy
  • Historical learning from your closed jobs

One job runs from the drawing set to the awarded subcontract with nothing retyped.

Firm

A whole estimating department, one shared history.

$549/month

or $5,490 a year — two months free

  • Everything in Pro
  • 20 seats
  • A separate workspace per project
  • Priority processing at tender time
  • Team rate sheets and cost books
  • Named contact and same-day support

Twenty estimators pricing against the same closed jobs rather than twenty private spreadsheets.

THE SAME PLAN OPENS THE BROWSER.

Whichever of these you pick, the programs it covers are also open on any machine with a browser — the office desktop, a site laptop, a borrowed one. Same projects, same rates, same numbers, no second subscription. Most firms subscribe in the browser and install the app afterwards; it makes no difference to what you pay.

See the plans in full →

Prices are the Canadian figures, charged in CAD, with tax worked out for your region at checkout. If you subscribe through the App Store or Play Store instead, Apple and Google bill it themselves in your own currency at their published rate for your territory, so the number in the store is the number you pay.

Who this is built for

PEOPLE WHO ARE
NEVER AT A DESK.

The browser is for the estimator with two monitors and a quiet hour. This is for everyone else on the job — the people whose office is a truck, whose desk is a tailgate, and who get asked the question standing in front of the thing.

Superintendents

The framer is asking about a detail now, not after you drive back. Open the sheet, read the finding, answer it.

Estimators on the road

Three site walks in a day and a bid due Friday. Price it between stops instead of at eleven at night.

Owners of small GCs

You are the estimator, the PM and the super. The phone is the only machine you have with you all day.

Anyone levelling quotes

Quotes arrive as photos of faxes and four-line emails. Forward them from the phone they arrived on.

The in-app assistant, offering questions like how to price a job, what levelling a bid does, and why a conflict confidence is low.

And it will explain itself

AN ASSISTANT THAT ONLY
ANSWERS TWO THINGS.

How to use the app, and questions about your own projects and documents. That is the whole scope, and it is enforced twice — the model is told to stay inside it, and separately is only ever given your account's workspace, so there is no question that reaches somebody else's estimate. Every answer names the sources it drew on, and it is built to say it cannot see something rather than invent a figure.

It is our own open-weights model, running on the same server as the engines. No third-party AI vendor is in the path of your tender documents.

Before you subscribe

WHAT THE PHONE
DOES NOT DO.

Your documents come to us.

The engines run on our server, so a document you audit is uploaded. It goes there and no further — your own isolated workspace, our own model, no third-party AI vendor, never used for training. If your contract requires that nothing leaves your own infrastructure, talk to us before subscribing.

It needs a signal.

Lists you have already opened stay readable offline and are stamped with when they were fetched. Running a new audit or repricing needs a connection. One bar is fine; a basement with none is not.

Big drawing sets are slow on cellular.

A 40 MB set is a 40 MB upload. It is streamed rather than held in memory, and the engine's real progress is shown as it works — but the physics of the upload are the physics of the upload.

It is not the full browser window.

Everything the browser can compute, the app can compute. What a phone cannot give you is four panes side by side. Deep catalogue editing and heavy rate-sheet work are still better on a laptop.

If none of that is a problem, the app is the fastest way into these programs. If any of it is, work in the browser — and take the app with you anyway, on the same plan.

iOS & Android · Q4 2026

IT IS BUILT.
IT IS NOT IN THE STORE YET.

Everything on this page is running today — that is why we could record it. What is left is review, store listings and the last of the Android build. It ships in Q4 2026, and if that slips this page changes and anyone waiting on it hears from us. It does not quietly become Q1.

Tell me when it lands → Buy the desktop and get it included The rest of the roadmap