PlanSwift vs On-Screen Takeoff.

PlanSwift vs On-Screen Takeoff: which one should you buy?

People often ask this as if one of the two were about to be retired. Neither is. ConstructConnect sells both, put its new AI into both this year, and keeps them apart because the people who use them work differently. The difference between PlanSwift and On-Screen Takeoff is less about measuring tools, which overlap heavily, and more about what happens to a quantity once you have it.

Pricing reviewed 13 November 2026, based on public documentation as of Nov 2026. PlanSwift, On-Screen Takeoff, Quick Bid and Takeoff BOOST are trademarks of their owner; this is an independent comparison and Pilars is not affiliated with ConstructConnect.

Side by side

One thing the table hides: neither program takes the drawings off your desk. Takeoff BOOST gives a first pass on some items, and an estimator still sets conditions, checks what the AI found and measures the rest.

Where PlanSwift is the better pick

PlanSwift treats a takeoff item as the start of an estimate. You drag an assembly onto a wall, and the line you trace becomes studs, plates, board, fasteners and labor, each with your own unit cost behind it. If your shop has spent years tuning those assemblies, that library is the product. PlanSwift also has strong Excel integration, so a shop whose bid lives in a workbook can keep that workbook.

The second reason is plain: you can see the price. Essential is $2,000 per year for one seat and includes Auto Count, Auto Scale, Auto Bookmark and two hours of one-on-one training. Core is $3,000 per year and adds Auto Takeoff, which is the full Takeoff BOOST set. ConstructConnect lists discounts for extra seats and volume pricing at ten or more licenses. A small builder can decide in an afternoon without a sales call.

The fit is strongest for residential and light commercial work, where one person measures and prices the whole job. Review sites are directional at best, but PlanSwift holds about 4.3 on Capterra across roughly 396 reviews, and the recurring complaints in third-party reviews are the learning curve, licensing glitches and support response.

Where On-Screen Takeoff is the better pick

On-Screen Takeoff has been the default in commercial interiors for a very long time, and the reason is its condition-based workflow. Multi-Condition Takeoff, which is patented, lets one pass along a partition book several conditions together, so the wall, the base and the ceiling perimeter do not each need their own trip around the room. Typical areas let you take off a repeated unit or floor once and apply it across the building. Overlays show what moved between addenda.

Take a drywall sub bidding a four-story medical office building with a dozen wall types. That estimator is not short of a pricing database; the shop already has Quick Bid or its own system. What that estimator needs is to get through 60 sheets of partitions without retracing anything. That is the job On-Screen Takeoff was built around, and it is why a commercial drywall sub will usually be happier there than in PlanSwift.

What you give up is transparency and scope. On-Screen Takeoff pricing is not public, so budgeting means a conversation with sales. And the program stops at quantities: the estimate is a second product. ConstructConnect cites a G2 rating of 4.4 out of 5 from 148 reviews.

Takeoff BOOST in both: what the AI changes

ConstructConnect announced Takeoff BOOST on 22 April 2026. It shipped first in On-Screen Takeoff and was announced as coming to PlanSwift, where it now appears on the pricing page. So the old shorthand that both are purely manual tools is out of date.

The details differ by product. In On-Screen Takeoff the published features are auto count of repeating items, Auto Name for sheets and wall-type detection. In PlanSwift the split is by tier: Auto Count, Auto Scale and Auto Bookmark come with Essential, and the fuller Auto Takeoff needs Core at $3,000. If the AI is why you are buying PlanSwift, price Core, not Essential.

Treat these as assists inside a manual program. They shorten the counting and setup. They do not read a specification, reconcile a door schedule against the plan, or hand you a finished quantity list. Use the 14-day trial on a job you have already bid, so you know the right answer before the software gives you one.

Cost for a real shop, and the Mac question

For three estimators, PlanSwift Essential comes to no more than $6,000 per year and Core to no more than $9,000, before whatever multi-seat discount you negotiate. There is no matching number for On-Screen Takeoff because none is published. Ask for the quote in writing with the seat count, the term and whether Quick Bid is included, then compare.

Third-party reporting says On-Screen Takeoff licensing is moving from license keys and server codes to email-and-password subscriptions during 2026. If you run older keyed licenses, ask your rep how that affects you. On-Screen Takeoff has not been discontinued.

Neither program runs natively on a Mac. PlanSwift's own system requirements say it cannot be installed directly on a Mac and needs Boot Camp or a Windows virtual machine. ConstructConnect's answer for Mac users is a third product, ConstructConnect Takeoff, which runs in a browser and carries the same BOOST AI; its pricing is also not public. More on this in PlanSwift for Mac and takeoff software for Mac.

If you would rather not click at all

Both programs assume an estimator drives. Pilars works the other way round: you upload the plan set and the spec book, the software reads them together, resolves the schedules and cross-sheet references, counts and measures, and returns quantities with a confidence score and a sheet reference on each one. The estimator's job becomes review.

It runs in a browser, so Windows versus Mac stops mattering. The price is $300 per user per month, billed monthly or annually, and each user covers one trade of their choice with unlimited takeoffs, so the bill does not move with the number of plan sets you run. That is $3,600 a year, more than a PlanSwift Essential or Core seat, so the case for Pilars is not the price. It is that the software does the first pass and the estimator corrects it: any detection can be opened and changed, and the totals update. Nineteen trades are covered, including drywall, painting, flooring and framing, and a shop that wants two of them needs two users. On counting accuracy we publish our own test, a duplex receptacle benchmark on a real power plan where the Pilars model scored 97.5%.

Pilars is not an assembly library or a Quick Bid replacement. If your pricing lives in PlanSwift assemblies, the sensible setup is to keep one seat for pricing and let the automated takeoff feed it. You can run your own set through the live demo first.

Questions estimators actually ask