The short answer: for a mid-sized fabrication shop — roughly 10 to 50 people, bidding 5 to 30 jobs a month — the best steel takeoff software is an AI-assisted tool that reads your PDF drawings, counts every member automatically, and lets an estimator verify the results before anything goes in a bid. That combination (AI detection plus human verification) is what gets you to 95-99% accuracy at a fraction of the manual takeoff time. General-purpose measurement tools like Bluebeam or PlanSwift still work, but at mid-sized volume the manual clicking becomes the bottleneck: an estimator spending 6 hours counting beams on a job you have a 1-in-5 chance of winning is the most expensive part of your bid process. The realistic 2026 shortlist is three or four steel-specific AI tools (SteelFlo, SketchDeck, Ferra, Beam AI), and the deciding factors are transparent pricing, whether you can test it on your own drawings before paying, and whether it covers the section standards you actually fabricate. Everything below is the detail: a 7-tool comparison table with pricing, what a takeoff seat should cost, and five tests to run before you sign anything.
What should a 10-50 person shop look for in takeoff software?
A mid-sized shop sits in an awkward spot. You bid too much volume for one estimator with a highlighter, but you're not big enough to staff a three-person estimating department that can absorb slow tools. That changes the priorities compared to what a one-man shop or a 300-person outfit cares about.
Here's what actually matters at this size, in order:
- Speed per takeoff, not features per screen. If your estimator does 15 takeoffs a month, cutting each from 5 hours to 1 saves 60 hours — most of a working month. Feature checklists don't win bids; turnaround does.
- Verifiable output. Automation you can't check is a liability on a lump-sum bid. The tool should show you exactly where every counted member sits on the drawing so a human can confirm or reject it. If it just spits out totals, walk away.
- Counts backed by a weight database. A takeoff that says "14 pieces of W12X26" is only useful if the tool also knows W12X26 weighs 26 lb/ft and can roll that into tonnage. Generic measurement tools leave you doing the weight math in a spreadsheet afterward.
- No per-seat trap. At 10-50 people you might have 1-3 people touching estimates. Per-seat pricing that looks cheap at one user gets ugly when the PM and the shop foreman also want to open a takeoff.
- A real trial on your drawings. Every vendor demos beautifully on their own clean sample PDF. Your drawings — scanned, rotated, half-raster, schedule keys everywhere — are the actual test. If a vendor won't let you run your own drawings before paying, that tells you something.
For a deeper walk through the whole category (what takeoff software does, where each tool starts and stops), see our steel takeoff software overview and the full 2026 software landscape comparison.
How do the main steel takeoff tools compare?
Seven tools come up in nearly every mid-sized shop's evaluation. Pricing below is list pricing or publicly stated positioning as of mid-2026 — always confirm current numbers with the vendor.
| Tool | Type | Steel-specific? | AI member detection | Pricing (as listed, mid-2026) | Best fit | |---|---|---|---|---|---| | SteelFlo | AI takeoff + BOM + pricing | Yes | Yes, with human verification step | $399/mo flat, 3 free takeoffs to test | Small and mid-sized structural shops that want speed with a review step | | SketchDeck (LIFT) | AI takeoff | Yes | Yes, includes connection/cope detection | Quote-based | Shops where connection-level detail drives the estimate | | Ferra | AI takeoff | Yes | Yes | Quote-based / early access | Shops willing to evaluate a newer entrant | | Beam AI | AI-assisted takeoff service | Structural + other trades | Yes, delivered as a service | Quote-based, per-project | Shops that want to outsource the takeoff rather than run a tool | | Bluebeam Revu | PDF markup + measurement | No | No — manual counting | ~$260-$440/user/yr | Shops already living in Bluebeam for markups and RFIs | | PlanSwift | On-screen digitizing | No | No — manual clicking | ~$1,749/yr per license | General estimating across trades, steel as a sideline | | Excel / templates | Spreadsheet | No | No | Cost of Office | Shops under ~5 bids/month with an experienced estimator |
A few honest notes on that table. Bluebeam and PlanSwift are good software — the problem for a mid-sized steel shop isn't quality, it's that neither one counts a beam for you. Every W-shape, every HSS, every angle is a manual click, and neither ships a steel weight database, so tonnage is a second manual pass. Beam AI has real traction (they cite 1,200+ customers) but operates more like a takeoff service than a tool you run yourself, which matters if bid turnaround is measured in hours. Among the steel-specific AI tools, SteelFlo is currently the only one with published pricing and a self-serve trial; SketchDeck and Ferra both require a sales conversation to learn what you'd pay.
If you're comparing the AI options specifically — what to ask vendors, how to read accuracy claims, where the failure modes hide — the AI steel estimating software buyer's guide goes tool-by-tool through that shortlist.
How much should steel takeoff software cost a mid-sized shop?
Work backwards from estimator time, because that's what you're buying back.
A mid-sized shop estimator costs roughly $75,000-$95,000 loaded. Call it $45/hour. If manual takeoffs run 4-6 hours each and you bid 15 jobs a month, that's 60-90 hours a month — $2,700-$4,000 of estimator time spent counting, before any pricing work starts.
| Approach | Cash cost / month | Estimator hours / month (15 bids) | Effective monthly cost | |---|---|---|---| | Excel + highlighter | ~$0 | 75+ | ~$3,400 in labor | | Bluebeam (2 seats) | ~$60 | 55-65 | ~$2,800 | | PlanSwift (1 license) | ~$145 | 50-60 | ~$2,600 | | AI takeoff w/ verification | $400-$800 | 15-25 | ~$1,300-$1,900 | | Outsourced takeoff service | $150-$500 per bid | ~5 (review only) | $2,300-$7,500 at 15 bids |
The pattern: anything in the $300-$800/month range for AI takeoff pays for itself if it saves even 10-15 estimator hours a month, and at mid-sized volume it typically saves several times that. The trap to avoid is quote-based pricing that scales per project or per sheet — at 15-30 bids a month, per-project pricing compounds fast, and you won't see it coming until the invoice. Flat monthly pricing is easier to budget and doesn't punish you for bidding more.
The other number to check: what happens on the jobs you don't win? You eat the takeoff cost on 70-85% of bids. A flat subscription spreads that; per-project outsourcing means you're paying full freight on every loss. Smaller shops running fewer bids do the same math with different inputs — we've broken that version down in steel estimating for small shops.
Do you need international standards support?
If every drawing you'll ever see is AISC wide-flange and HSS, maybe not. But two situations make multi-standard coverage matter more than shops expect:
- Imported or foreign-engineered projects. Equipment skids, PEB packages, and industrial work engineered overseas regularly arrive with EN (European), IS (Indian), or GB (Chinese) sections on the drawings. If your software's database doesn't know what an HEA200 or ISMB300 weighs, you're back to manual lookups on exactly the jobs with the least familiar material.
- Cold-formed steel. Mixed structural/CFS jobs (mezzanines, infill framing) need SSMA stud and track sections in the database, or half the BOM gets hand-built.
Coverage varies widely: Bluebeam and PlanSwift carry no steel database at all, most AI tools cover AISC only, and SteelFlo covers 4,500+ sections across six standards (AISC, BS/IS, AS/NZS, EN, GB, CFS/SSMA) — currently the widest coverage in the category. If even 5% of your work involves non-AISC sections, put standards coverage on your evaluation checklist; it's nearly impossible to retrofit.
How do you evaluate takeoff software on your own drawings?
Don't buy off a demo. Run these five tests on your own PDFs during the trial — they take an afternoon and expose the differences that never show up in a sales call.
- The scanned-drawing test. Feed it your ugliest job: a scanned, slightly rotated raster PDF. Vector-only tools fall over here, and a lot of your incoming drawings aren't clean vector.
- The count audit. Pick one sheet, count every member by hand (yes, all of them), then compare against the tool's output. You want to see both the miss rate and the false-positive rate — a tool that finds 105 members when there are 100 is worse than one that finds 97, because phantom steel inflates your bid.
- The schedule test. Give it a drawing where members are labeled by schedule key (B1, C2) with sizes in a table. Many tools count only direct labels and silently miss schedule-referenced members — often the majority of the job.
- The verification test. Deliberately find one wrong detection and fix it. How long did that take? Could you see where on the drawing the count came from? If reviewing the AI's work is slower than doing the takeoff yourself, the accuracy number is irrelevant.
- The export test. Get the BOM out and into your pricing workflow (or the tool's own pricing). Count the manual steps between "takeoff done" and "bid number." That gap is where tools quietly give back the time they saved.
Run all five on two or three tools and the decision usually makes itself.
Is there free steel takeoff software?
Genuinely free, steel-specific takeoff software doesn't exist in any form you'd trust a bid to. Your real free options are Excel templates plus manual counting (free in cash, expensive in hours) and free trials of paid tools. On the trial front, look for one that doesn't require a sales call — of the tools in the table above, only one offers a self-serve trial (3 free takeoffs) with no demo appointment, which is enough to run every test in the section above on your own drawings before spending anything. Treat "free" trials that require a sales call and a credit card as what they are — a sales process, not a trial.
Should a mid-sized shop outsource takeoffs instead?
Outsourcing (takeoff services, typically $150-$500+ per bid depending on job size) makes sense in two cases: overflow during bid surges, and shops with no estimating capacity at all. It makes less sense as your standing process at mid-sized volume, for three reasons: per-bid cost compounds across the 70-85% of bids you lose; turnaround is measured in days, not hours, which costs you late-drop opportunities; and you still have to review someone else's takeoff before you bid it — without the drawing-level visibility that makes review fast in a good tool. Most mid-sized shops land on software as the default with an outsourced service as surge capacity, not the other way around.
If an AI takeoff with a built-in verification step sounds like the right shape for your shop, see SteelFlo's pricing — a flat, transparent $399/month for SteelFlo Pro, with 3 free takeoffs so you can run every test in this guide on your own drawings before paying a dollar.