How to make a Revit plugin for tagging, and what it takes to keep it working

The working method

Serious Revit plugins are built in .NET with C#. This is the route that gives you full control over the API, performance, and a real interface.

1. Create a class library that targets the .NET version of the Revit release you support, and reference the Revit API assemblies.
2. Implement IExternalCommand and register it with an .addin manifest so Revit loads it.
3. Collect the elements in the active view with a FilteredElementCollector that is scoped to the view.
4. Open a Transaction, and for each element call IndependentTag.Create with the view, the element's reference, and a position. The overloads differ between Revit versions.
5. Commit the transaction, and roll it back if anything fails.

That produces tags. It does not produce good tags.

Where it breaks

Placing a tag is easy. Placing it well is the work. A tag at the element's midpoint lands on ducts, fittings, and other tags. To avoid that, you need the 2D footprint of every element as it is drawn in the view, the real size of each tag, and a search for free space. Much of the API is undocumented, and the explanations are in third-party blogs.

Every view changes the logic. Plans, sections, rotated views, cropped views, linked models, and fabrication parts each behave differently. Something that works for one element breaks as soon as the view changes. There are too many edge cases to cover from a single test project.

Your firm's standard. Tags inside or outside the duct, above or below, leader angles, and where size and elevation sit are all firm rules. You have to build and maintain every one of them.

The code rots. Race conditions, hidden state, swallowed errors, and misleading fallbacks are the usual mistakes. They turn into silent failures, where a tag is missing and nothing tells you. This is the complexity wall that quick plugins hit.

Revit changes every year. Revit ships a new version annually and several update releases in between. Revit 2025 moved to .NET 8 and required every add-in to be recompiled, with API changes across several areas. Revit 2027 moves to .NET 10. Tag-creation overloads and orientation handling also differ by version. Staying current, and moving to newer and faster APIs when they appear, is a cost that never goes away.

Crashes and lost work. Plugins run inside Revit, so an unhandled exception, a call from the wrong thread, or an unfinished transaction can take Revit down. Revit has no native autosave, so a crash can wipe out two or three hours of work. You also have to build the undo yourself, as one undo step per operation, or a bad run cannot be reversed.

Performance. Large views with thousands of elements need careful use of the API, or the run freezes Revit.

What it costs

If your firm has a dedicated BIM developer and tagging is a narrow task you own, building makes sense. Otherwise the add-in becomes a second job for whoever wrote it. RayCAD is the maintained version of this: a .NET plugin that handles placement, firm standards, undo, crash safety, and Revit 2021 to 2027. Read the full build vs buy comparison and see how RayCAD runs inside Revit.

Try RayCAD