Structural steel detailing software: Tekla, SDS/2 and the alternatives.

What steel detailing software has to do

Steel detailing software has one underlying job regardless of brand: turn a structural model into a shop that can cut, drill and weld without guessing. That breaks into a handful of capability categories worth judging separately rather than as one score.

• Connection automation — how much of a standard connection the software detects and details itself, against how much a detailer builds by hand.
• Drawing production — how directly the software turns a checked model into erection and single-part drawings that need minimal manual cleanup.
• Fabrication output — whether NC1/DSTV, DXF and management-system exports come out complete and correct the first time.
• Marking and numbering — how the software assigns piece and assembly marks, and how reliably it keeps them straight when the design changes after drawings are already out.
• Concurrent working — how many detailers can be in the same model at once, and what happens to their work when it is merged.
• Interoperability — how cleanly the model exchanges as IFC with architecture, services and analysis software outside the steel package.
• Learning curve and licensing model — how quickly a team becomes productive and how a business plans for it.

No package leads in every category, which is why the comparison below is organized by what each tool is actually good at rather than a single ranking.

One thing a capability list tends to hide: the primary frame is rarely where the hours go. Stairs, handrails, ladders, walkways, gratings, embeds and the rest of the secondary and miscellaneous steel carry far more individual parts than the columns and beams do, and they are where a component library, a custom-component editor and a numbering scheme are genuinely tested. A package that details a portal frame elegantly and then leaves you drawing every handrail return by hand has not saved the programme anything.

Tekla Structures

Tekla Structures, developed by Trimble, is the most broadly adopted detailing environment worldwide, and its strongest argument is reach: fabricators on every continent already work from Tekla output, and the software handles steel, precast and reinforced concrete inside one model rather than requiring separate packages. Its component system covers an extensive library of standard connections and lets a detailer build custom components for anything the library does not already cover, which suits complex or unusual geometry.

Tekla's open API is a genuine differentiator — it is the platform most third-party plugins, connection-design add-ins and fabrication-management integrations are built against first, and its IFC support is generally considered among the strongest in the category. Model size is rarely a limiting factor either — Tekla is a common choice on the largest and most geometrically complex programmes precisely because it holds a big, multi-material model without the performance problems a narrower tool can run into. The trade-off is a steeper learning curve than some alternatives and a model that rewards a well-organized template; an undisciplined Tekla model is harder to untangle than the equivalent in a narrower tool.

SDS/2

SDS/2, developed by Design Data, is built around connection design rather than connection placement — it calculates connection capacity against the governing code as it details, which shifts work that is manual in some other packages into the software itself. That makes it a strong fit where AISC-governed steel dominates the programme and connection engineering time is the bottleneck.

Its drawing automation is correspondingly strong: because the software already knows the engineering behind a connection, producing a correct detail drawing from it is a smaller step. SDS/2 is narrower than Tekla in scope — it is a steel-focused tool first — and its ecosystem of third-party integrations is smaller, but for a fabricator whose work is mostly structural steel under a single code, the automated connection design is the reason it gets chosen.

Advance Steel and the alternatives

Advance Steel, developed by Autodesk, sits naturally inside a Revit and AutoCAD environment, which is its real advantage: teams already standardized on Autodesk tools get steel detailing without introducing a second modeling philosophy, and DWG-native drawing production is a direct strength of being built on AutoCAD. Its connection automation and large-model performance are generally regarded as narrower than Tekla's or SDS/2's, which makes it a stronger fit for less geometrically complex steel than for a stadium roof or a heavily connected industrial frame.

Regional and specialist detailing tools exist beyond the three main platforms too, built for particular fabrication contexts or national markets. The honest answer for any of them is the same one that applies to the majors: the software matters less than whether its output matches what your fabricator's shop floor and management system actually consume.

Marking, numbering and what the model checks for you

Piece and assembly marks are the thread tying a drawing to a part on the shop floor, and automatic numbering is one of the least glamorous and most consequential features in any detailing package. All three of the main tools assign marks from the model rather than by hand; what separates them is behavior on revision. When a plate thickness changes on one of forty otherwise identical assemblies, the question is whether the software gives that one assembly a new mark and leaves the other thirty-nine alone, or renumbers the lot and invalidates a set of drawings already out for fabrication. Numbering settings decide that, and getting them wrong is expensive in a way that does not surface until the second issue.

Model checking is the other half. Clash detection inside the steel package catches the interferences the frame creates on its own — a brace working line running through a beam flange, a bolt group with no wrench access, a haunch fouling a purlin. Tekla and SDS/2 both carry this natively, and SDS/2 ties it to the connection engine so a connection that cannot physically be built is flagged as it is created rather than at a review weeks later. Neither replaces coordination against the federated model, where the clashes that actually stop a site are usually steel against services rather than steel against steel.

What no clash engine looks at is the rest of it: marks reconciled against the bill of material, fabrication data against the drawing it claims to describe, and the model against the standard details it was supposed to follow. Those are the checks we automate on top of whichever package produced the model, and the record of them is issued with the package rather than kept internally — the discipline set out in our note on quality control in engineering drawings.

Several detailers, one model

Any steel package of real size is detailed by more than one person, and how the software handles that shapes the programme more than most feature comparisons admit. Tekla Structures uses a model-sharing approach: each detailer works in a local copy and reads and writes changes back to a shared version, with a history that can be stepped through and, where necessary, unwound. SDS/2 takes the more direct route of several users in one live model, with edits appearing to everyone as they happen.

Both work, and they fail differently. Shared-copy working tolerates a poor connection and an offline day, but it needs the discipline to write in often enough that two people are not detailing the same bay from different starting points. Live multi-user working removes that risk and replaces it with a dependency on the network and on people staying out of each other's areas. Either way, the thing worth agreeing before modeling starts is the split — who owns which zone, which phase or which erection sequence — because a model divided along a sensible work breakdown stays workable at four detailers, and one divided along nothing stops being workable at two.

Codes, profile databases and regional standards

A detailing package carries the code it was built around in more places than the code checks. The profile catalogue, the bolt and anchor libraries, the weld symbol conventions, the default drawing presentation and the tolerances a connection is designed to all reflect a region of origin. SDS/2's connection engine is written around AISC practice, and its automated design output is at its most persuasive on North American steel. Tekla ships regional environments — profile and bolt databases, component behavior and drawing settings per market — which is a large part of why it moves between Eurocode, British Standards and Australian Standards markets as readily as it does.

Two practical consequences follow. A package configured for the wrong region will happily produce a correct-looking drawing carrying the wrong section designations and the wrong bolt grades, and it will do it silently. And a national annex is not a software setting: where a Eurocode job is delivered in a tool whose defaults are American, the connection design behind it has to be checked against the annex in force by someone who knows it, whatever the software believed. That is one of the reasons the check and the signature sit outside the modeling tool in our process rather than inside it.

How to choose: the questions that decide it

The question worth asking first is not which software is best, but what the fabricator's shop already runs on — the fabrication format a package has to produce is usually the least negotiable requirement on the list. After that: how much of the connection design is delegated to the detailer versus already fixed by the engineer of record, since that shifts value toward SDS/2's connection-design engine or toward Tekla's component flexibility depending on the answer.

Also worth asking: whether precast or reinforced concrete needs to live in the same model as the steel; how the project exchanges data with architecture and services, since IFC quality matters more on a heavily coordinated job than on a standalone shed; and, plainly, what your own team or delivery partner already has deep experience in, since fluency in a tool changes output quality more than a feature list does. Licensing model is worth factoring in as well — seat-based and programme-based arrangements suit different team shapes, and it is a genuine cost variable even though we do not publish prices here, since they change and are negotiated per account.

Delivering in whichever tool your project uses

We deliver in Tekla Structures, SDS/2 and Advance Steel, set up against whichever your fabricator already runs on rather than asking a shop to adapt to ours. The AI-driven part of the work — framing, connection instances against your standard details, drawing creation, marking and the bill of material — is configured to the platform in use, and the automated checks run against the model in that same environment before a senior engineer reviews and signs the package.

Where a programme spans more than one platform, or needs to exchange with architecture or services software outside the steel package, IFC is issued alongside the native model and the two are checked against each other before issue, so a change in one is never quietly missing from the other.

Where your organization would rather the work happened inside its own licensed environment, that is how delivery is set up, so the model stays where your data policy says it should sit. The sectors this comes up in most are the ones with the heaviest steel packages and the tightest coordination — bridges, airports, metro, power plants and industrial buildings — and in every one of them the tool matters less than whether the numbering, the fabrication data and the check record hold up across a package issued in stages.