Mechanical estimating,
done by AI.

Mechanical is really four trades

When people say "mechanical," they mean a stack of scopes that don't behave the same way. HVAC piping — chilled water, hot water, refrigerant, condensate — runs as risers and mains with diameters and fitting counts that drive both material and labor. Ductwork is measured as runs and sizes, then converted to area and weight by gauge. Plumbing brings its own waste, vent, and domestic-water systems with their own fixture counts. And sitting over all of it is the equipment: air handlers, pumps, boilers, chillers, VAV boxes, every one carried on a schedule and a tag.

A takeoff tool built for one of those — say, fixture counting — falls down on the others. Mechanical estimating software has to handle the whole stack, because a real mechanical bid touches all four on the same set of drawings. If you only bid one scope today, our HVAC takeoff and plumbing takeoff pages go deeper on each.

What Pilars reads on a mechanical set

Where mechanical bids leak

Every estimator who's bid mechanical knows the three places money quietly disappears. They're not exotic — they're routine, which is exactly why they slip through.

• Duct weight conversion. You measure area and gauge, then convert to pounds. Do that segment by segment under deadline and small errors compound across a building's worth of duct. Consistent measurement off the drawings is the only real fix — guess the area and the weight is wrong before you start.
• Fitting counts. Elbows, tees, transitions, and reducers carry real labor and material, and they're the first thing skipped when a manual takeoff runs late. Undercount fittings and your piping labor is light from the moment you submit.
• Schedule-versus-spec drift. The equipment schedule says one thing, the spec section says another, and the plan tag references a third. When they disagree, somebody buys the wrong unit. Reading the schedule and the spec together — and tying both to the tag — is how you catch the conflict during the bid instead of during buyout.

How a mechanical takeoff actually goes

Walk through a typical mechanical set and you can see why a single-purpose tool struggles. You open the HVAC plans and the duct is layered over the ceiling grid, the architectural background, and the electrical — and you need to pull just the duct, by size, without dragging in everything around it. Then the piping sheets: chilled water and hot water mains threading the same corridors, condensate running off every unit, refrigerant lines to the rooftop equipment. Each system has its own diameters, its own fitting density, its own labor rate. Miss that a 6-inch main steps down to 4-inch halfway across the building and your material and labor are both off.

Then you turn to the schedules and the spec book, which is where the real scope lives. The equipment schedule lists every air handler, pump, and VAV box with capacities and tags; the plumbing fixture schedule does the same for sinks, water closets, and floor drains; the spec sections dictate insulation thickness, valve types, hanger spacing, and the test-and-balance scope that never shows on a plan. A mechanical estimator is constantly cross-referencing all of it — plan to schedule to spec — and the errors come from that cross-referencing breaking down under deadline. Pilars reads those documents together and keeps the tags, quantities, and specs tied to each other, so the cross-reference doesn't fall apart at 11pm the night before bid day. That's the part of the job most tools simply ignore, and it's the part where mechanical bids are won or lost.

Review, don't trace

The shift with AI takeoff is simple but it changes the whole day: instead of starting from a blank set and tracing every duct run and counting every fixture by hand, you start from quantities Pilars has already produced and spend your time checking them. That's where an estimator's judgment actually earns its keep — confirming the questionable run, sanity-checking the count against your gut, deciding how to carry the ambiguous detail — instead of being spent on the mechanical labor of measuring. On a multi-scope mechanical set, that's the difference between bidding two jobs this week and bidding five. You can see exactly how the read works on the how it works page, or watch it on the live demo.

An honest note on scope

We'd rather tell you the edges than oversell them. Pilars automates the measurable, drawing-and-schedule part of mechanical takeoff — runs, sizes, diameters, fitting counts, equipment tags, fixture counts, and the spec scope tied to them. What it doesn't do is replace your means-and-methods judgment: how you'll route in a tight ceiling, what the field crew will actually hit, how you price the risk on a vague detail, or which assemblies and labor rates your shop carries. Genuinely ambiguous drawings still need an estimator's read, and we flag those rather than guess. The goal isn't to remove you from the bid — it's to take the tracing off your plate so the hours go to the parts that need a human. Where a detail is unclear, you'll see it called out, not silently assumed.

What it costs

Mechanical takeoff with Pilars is $100 per trade per plan, with no per-seat fees. Bid mechanical as one trade and your whole team — estimators, reviewers, the seasonal hand you bring on for a busy spring — works under that one flat number. No license counting, no per-project credits, no plugins to unlock the schedule reader. Compared with per-seat platforms that charge each estimator separately, a mechanical shop with more than one bidder usually comes out well ahead. If you're weighing options, our pricing comparison lays the whole market out side by side.

Mechanical estimators ask