Best AI App for Carpenters: Measuring, Quoting and Invoicing Compared

No single app wins all three jobs. Solo carpenters do well with a tape or laser for measuring, a cut-list optimiser for sheet goods and a quoting app with AI drafting: ServiceM8 is free for one user up to 30 jobs a month, and Joist starts at about $10. CountBricks adds plan takeoff with AI estimates at $30 per user.

The AI in these apps drafts text and line items. It does not measure to joinery tolerance, so the numbers you feed in decide whether a quote makes money. The way to choose is to run one real job through each candidate. The wardrobe below gives you the inputs, the right answers and a scoring sheet. I read vendor pages for this tutorial; I did not bench-test the apps, so the test is yours to run.

Follow me on Instagram@sagnikteaches

The wardrobe job: one test for all three tasks

An alcove about 2.4 m wide and 2.3 m high gets a built-in wardrobe. The carcase is 18 mm plywood, assembled in the alcove, with three bays and two shelves in each. The customer supplies the doors, someone else paints, and you take a 30% deposit. Every figure below is illustrative, prices included, so swap in your own.

Connect on LinkedInSagnik Bhattacharya

These are the right answers to check each app against:

Subscribe on YouTube@sagnikteaches

• Measuring: widths of 2,412, 2,405 and 2,398 mm at top, middle and bottom, and heights of 2,304, 2,300 and 2,296 mm. The walls lean, so the carcase is built to the smallest of each less a fitting allowance: 2,390 wide and 2,280 high.
• Cut list: 12 parts that need exactly 4 sheets of 2,440 × 1,220 mm (the 8 ft by 4 ft sheet), with about 18.4 m of edge banding on the front edges.
• Quote: cost $1,666, price $2,220 at a 25% margin.
• Invoices: a deposit of $666, a $180 change order for two extra shelves, then a final invoice of $1,734 that shows the deposit received.

Score each tool 0, 1 or 2 on the checks that apply to it: a cut-list tool on check 2, a quoting-and-invoicing app on the other six, for 12 points. Also note the minutes from blank screen to sent final invoice. The score decides. On a tie, prefer the app whose weak checks only you see, then the faster one.

Measuring: what each app can and cannot give a carpenter

Apple describes the readings from its built-in Measure app as approximate, and it works best at roughly half a metre to three metres. That suits quantities and sketches. It does not suit a wardrobe that has to slide into an alcove. In the wardrobe job the tape gives 2,412, 2,405 and 2,398 mm. A phone reading of 2,430 mm would be 32 mm more than the narrowest point, and a carcase built to it would not go in. Build to 2,398 less a fitting allowance, 2,390 here, and scribe the gap. The site measure checklist tutorial shows how to capture notes and photos so these readings survive until the workshop.

Put a price on the miss. A carcase built 32 mm too wide has to come apart, and the top, bottom and shelves are re-made: about two sheets ($136 at $68 a sheet) plus perhaps six hours of labour ($330), or $466 on a $2,220 job, 21% of the price. A tape costs a few dollars. Whatever the app, a routine like this keeps the numbers honest:

1. Read the width at three heights and the height at three positions, and note the smallest of each.
2. Check floor, ceiling and wall with a level and note any lean.
3. Photograph each reading with the tape or laser in frame and the number legible, so the file proves it.
4. Enter millimetres, not centimetres rounded, and make sure the app keeps what you typed.
5. If you assemble a carcase before it goes in, check it can stand up: the square root of (height² + depth²) must be less than the lowest ceiling reading. Here that is the square root of (2,280² + 560²), about 2,348 mm, against a 2,296 mm ceiling, which is why this one is assembled in the alcove.

The paid scanning apps price themselves for different users:

• Polycam: Free; Basic $150 a year or $30 a month; Business $400 a year per user. Every tier has a ruler tool, but the plan descriptions put automated measurements and 2D and 3D floor plans on Business, and the page's only accuracy line is that more images give more detailed and accurate models.
• Magicplan: priced per project, $25, $30 or $40 depending on a 24-month, 12-month or month-to-month commitment, in blocks starting at 10 projects a month, so $250 to $400 a month, with $40 for each extra project. New users get two free projects. It lists laser-assisted measurements, LiDAR scanning on iOS Pro devices and built-in estimating, and describes no AI features beyond a help prompt and a link for connecting Claude or ChatGPT.

A carpenter finishing three or four measured jobs a month would pay for ten on Magicplan, with unused projects rolling over. These tools are built for people who scan many rooms, so for one-off built-ins a laser measure and a notebook still win on cost.

Cut lists: the sheet-count check

The 12 parts are two sides at 2,280 × 560 mm, two dividers at 2,244 × 560 (they sit between top and bottom), a top and bottom at 2,354 × 560 (between the sides), and six shelves at 772 × 540. Their combined area is 10.20 square metres. One sheet is 2.44 × 1.22, or 2.977 square metres. Divide and you get 3.43, so the minimum is four sheets. A hand layout confirms four: two long parts fit side by side across the 1,220 mm width (560 + 3 mm kerf + 560 = 1,123), so six long parts use three sheets, and the six shelves fit three along (3 × 772 plus two kerfs is 2,322) by two across on the fourth.

That is a test any tool must pass. An answer of three sheets is wrong; an answer of six has not packed well. At $68 a sheet, the gap between four and six sheets is $136 on one job.

Edge banding is the other quantity apps get wrong. Only the front edges show in a built-in, so the sum is six shelves at 772 mm, two sides at 2,280, two dividers at 2,244 and a top and bottom at 2,354: 4,632 + 4,560 + 4,488 + 4,708 = 18,388 mm, or 18.4 m. Add 10% for waste and you need about 20.2 m. A tool that quotes 12 m has counted the wrong edges, and one that says 50 m or more has banded every edge.

• The cut-list site at cutlistoptimizer.com: panels and stock tabs, a cut-thickness (kerf) setting, labels, edge banding and orientation options, and results that report layouts, used and wasted area, number of cuts and any panels it was unable to fit. It exports PDF or an image. The page shows a three-day unlimited pass for $4.90 and mentions Bronze, Silver and Gold subscriptions without prices, so check the limits before relying on it.
• TimberLab: an App Store tool that generates a 3D model and a cut list from a typed description or a photo. It is free to download, with TimberLab Unlimited at $49.99 a year or $8.99 a week. The listing promises an exact "wood block" cut list to take to a mitre saw, never mentions sheet goods and carries no warning to check the output, so test it on a sheet-goods job before you trust it.
• A chat assistant: good for naming the parts of a wardrobe, poor at packing them. Use the area test on whatever it returns.

The cutting list software against AI tutorial goes deeper on reliability, so I keep this section to the one check that catches most errors.

Quoting: price the same job in five apps

Start with the sum that apps disagree about. Cost is $456 of materials (four sheets at $68 is $272, edging and fixings $64, door hardware $120) plus 22 hours at $55, which is $1,210. That makes $1,666. A 25% margin means price = cost ÷ 0.75 = $2,221, which I round to $2,220. A 25% markup gives $1,666 × 1.25 = $2,082.50, which is only a 20% margin. The difference is $139 on one wardrobe. Apps label the percentage field differently, and for the same price the markup figure is always the higher of the two, so know which one your app asks for.

Feed each app the same note and compare the draft. Here is a prompt for a chat assistant, with the sort of draft that needs editing:

Draft quote lines for this job from the notes only. Do not add work that is not in the notes. Notes: built-in wardrobe in an alcove, 2.4 m wide, 2.3 m high, three bays with two fixed shelves each, 18 mm ply carcase, customer supplies bi-fold doors, paint by others. Output numbered lines with description and quantity, no prices.

1. Supply and fit 18 mm plywood carcase, three bays, to alcove 2.4 m wide by 2.3 m high. 2. Six adjustable shelves, two per bay. 3. Fit customer-supplied bi-fold doors. 4. Edge banding to exposed edges. 5. Painting and finishing of carcase. 6. Removal of waste.

Three lines would not go out. "Adjustable" contradicts the notes, which say fixed. The painting line is work someone else is doing. Waste removal may be fair, but only if you have priced it. Count invented items in every app's draft; an app that adds three on a six-line job needs a firmer prompt or a different habit. The tutorial on testing day rates against job history helps you decide whether the $55 hourly figure is right in the first place.

A second realistic miss is hours. If the notes say nothing about access, floors or walls, an AI draft can supply a generic labour figure. Suppose it proposes 12 hours where your history says 22 for a built-in like this. At $55 an hour that is a $550 hole that no markup rescues. Feed hours from your own past jobs, or tell the assistant to leave labour blank.

Invoicing: a deposit, a final bill and a change order

With a quote of $2,220 and a 30% deposit, the deposit invoice is $666 and the final is $1,554. Now the customer asks for two extra shelves at $90 each, $180 in all. The final invoice becomes $1,734 (the new $2,400 total less the $666 received), and it must show that deposit as a line, not just a smaller number. Run that sequence in every app, including the change order, because that is where invoicing tools differ.

• ServiceM8: its plan table lists deposits and partial or progress invoices, and a progress-invoicing add-on from Starter. QuickBooks and Xero sync are listed on every plan.
• Tradify: progress invoicing and automatic reminders arrive on Pro at $51 per user. Xero and QuickBooks Online sync are on every plan.
• Joist: its comparison table lists deposit and milestone payments, though the page I read does not show which plan has them, and Elite adds change orders. Payment processing fees apply and no rates are listed.
• Jobber: customers can approve a quote and pay a deposit online, and quote follow-ups need Connect.
• CountBricks: keeps estimates, proposals, changes and invoices in one connected workflow, but its pages name no accounting integration, so ask how a paid invoice reaches your books before you rely on it.

The stage-payment wording is worth reading before you send the first one, and the builder invoice template tutorial has sample requests. Check three things on the invoice itself: the deposit shows as received, the change order is its own line, and the total matches your spreadsheet to the cent.

What the AI in each app actually does

"AI" on a pricing page covers very different things. This is what each page names, and what is left to you.

The pattern is plain. The AI sits on text, line items and counting. None of the pages promises a measurement good enough to cut to. That makes your test the only evidence that counts.

What general assistants add, and where they trip

A chat assistant is the cheapest tool here: ChatGPT Plus and Claude Pro are each about $20 a month and Google AI Pro is $19.99, all list prices that change, and free tiers exist with lower limits. It earns its place in three ways. It turns rough notes into a quote description. It rewrites an invoice email or a payment reminder in your voice. And it checks arithmetic you give it. Try the last one with the wardrobe costs:

Cost is $1,666. Give me the selling price for a 25% margin and for a 25% markup, show the working, and say what margin the markup version actually gives.

A good answer gives $2,221 and $2,082.50, and a 20% margin for the second. If the assistant mixes the two up, that is useful to know too. It trips on packing sheets, on measuring from photos with no scale, and on prices it has not been given. Keep it away from your prices: it will fill a gap with a plausible number that is yours to supply. The joinery AI app checklist lists what to check before paying for anything newer.

Running the wardrobe test in one sitting

Give each quoting app one uninterrupted hour. Use a fake customer called "[first name] Test" with your own email and phone, so every quote and invoice comes back to you, and run the steps in this order:

1. Start the clock. Add five price-list items: 18 mm ply at $68 a sheet, edging and fixings at $64 a job, door hardware at $120, labour at $55 an hour, and an extra shelf at $90.
2. Type the dimensions, 2,390 by 2,280 mm, into the quote description, save, and reopen it to see whether the app kept them (check 1).
3. Paste the job notes into the app's AI drafting, or use the prompt above, and count invented lines before you delete them (check 4).
4. Build the quote from four sheets, edging and fixings, hardware and 22 hours, set a 25% margin, and see whether the total reaches $2,220 with cost and margin on screen (checks 3 and 5).
5. Send the quote to yourself, open it on your phone and accept it, noting whether the app asks for the deposit at acceptance.
6. Raise the 30% deposit invoice for $666 and mark it paid (check 6).
7. Add the change order for two extra shelves, raise the final invoice, stop the clock and compare the balance with $1,734 (check 7).

The cut-list tool gets a shorter run. Type the 12 parts, set a 3 mm kerf and a 2,440 × 1,220 mm sheet, band the front edges only, and read off three numbers: 4 sheets, about 18.4 m of banding, and no part listed as unable to fit. For a measuring app, measure one real opening three times with the app and three times with a tape. If the largest gap is bigger than your fitting allowance, 8 mm in this job, the app is for sketches and quantities, not for cutting.

Here is a filled-in sheet for two imaginary quoting apps, to show how one reads:

The totals tie, so the tie-break applies, and it does not simply hand the win to the faster app. App B's weak checks, a quote screen without the margin and one invented line to delete, are both caught by you before anything leaves the office. App A's 1 on check 7 means every final invoice hides the deposit inside a smaller total, which a customer reads as a mistake and rings about. App B is the keeper here, and its extra 13 minutes should shrink once the price list is set up. Run the second app the next day rather than straight after the first, or practice from the first run makes the second look quicker than it is.

Cost and learning curve, side by side

Every quoting app needs the same setup before it saves time: a price list of your 30 to 50 most common items, your labour rate, a margin rule, a quote template and your payment settings. Plan an afternoon for that, and treat the hour count as my estimate, not a vendor figure. The learning curve is easiest to see with a stopwatch. Build the wardrobe quote on day one and again on day five. If the second run is not at least a third faster, the app is fighting you.

The workshop line leaves out the cut-list subscription, because the page I read shows no price for it. Add the figure before you compare, and note that Tradify's page does not name carpenters, so check its fit with your own quotes in the 14-day trial.

Which two apps to trial first, by how you quote

If you quote from site visits, which is most carpenters, run the wardrobe test on ServiceM8's Free plan and one paid alternative such as Joist or Tradify, and keep whichever wins on the score sheet. If you bid from drawings, add CountBricks to the test and give it a real plan set as well as the wardrobe. If you only want one change this month, put the sheet-count check and the margin sum into your routine, because those two catch more money than any app. Nothing here needs more than two trials, and each is free or short, so the cost of choosing wrongly is a week, not a year.

Questions carpenters ask about AI apps

Can AI measure a room accurately enough for built-in joinery?

Use it for planning, not for final dimensions. Apple describes its Measure readings as approximate, and a built-in unit has to fit the narrowest point of a wall that may lean. Take final sizes with a tape or laser at the top, middle and bottom, and build to the smallest reading less a fitting allowance.

Can ChatGPT or Claude make a cut list?

They can list parts from a description, but packing parts onto sheets is the job of a cut-list optimiser. Check any answer with an area test: total part area divided by sheet area gives the minimum sheets. If the answer is below that figure it is wrong, and if it is well above it, the layout is wasteful.

Is there a free way to quote and invoice as a carpenter?

ServiceM8's Free plan covers one user and 30 jobs a month, with quoting, invoicing and an AI writing helper capped at 10 assists a day. Joist ended its free plan; its cheapest plan is about $10 to $12 a month, with a 14-day trial and no card needed. Free tiers change often, so confirm on the vendor's page before you build your routine around one.

How many apps does a carpenter really need?

Three jobs, so usually two or three tools: something to measure with (often a tape or laser), a cut-list tool if you work with sheet goods, and a quoting-and-invoicing app. A general chat assistant can sit across all three for wording and arithmetic checks, though it should not be the system of record.

Further reads

• Cabinet Shop Software: Quotes, Cut Lists and Orders, and What AI Adds — Workshop software for quotes, cut lists and orders.
• AI App for Tradies: ServiceM8, Tradify or Fergus by Business Size — Choosing ServiceM8, Tradify or Fergus by business size.
• Blinds Quoting Software: Measure-and-Quote Tools and What AI Adds — Five blinds quoting tools compared on price lists, on-site measuring and cost, plus what their AI really does and a 20-cell audit to run before you trust it.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: ServiceM8, Jobber, Tradify, Joist, CountBricks, Polycam and Magicplan pricing and feature pages, the cutlistoptimizer.com tool page, the TimberLab App Store listing and Apple's iPhone User Guide page on Measure, all read in October 2026.