Automation that survives the next revision.
A macro encodes one sequence and breaks when the tree changes. CADABRA reads the model and resolves what you asked for against what is actually there.
Why the macro library stops growing.
The reason is structural: a macro addresses features by name and position, so it works until somebody renames a feature, reorders the tree, or models the next part slightly differently. When it breaks, someone has to open the code.
The cost is not writing the first macro. It is that every variation needs another one, and each one has to be maintained.
Macro, API script, or copilot.
| Recorded macro | API script | CADABRA | |
|---|---|---|---|
| Who writes it | Any engineer, once | A developer | Nobody, you describe the outcome |
| Handles a renamed feature | No | Only if coded for it | Yes, it resolves against the model |
| Handles a part it has not seen | No | Rarely | Yes |
| New variation costs | A new macro | A code change | A different sentence |
| Maintenance | Silent breakage | Ongoing developer time | None to maintain |
| Best used for | A fixed, repeated click sequence | Deep, stable integrations | Everything that varies |
The work worth automating first.
Where customers start.
Batch drawings
Family-wide edits
BOMs from the assembly
Component lookup
Legacy conversion
Error triage
Keep the macros that work.
A macro that runs a fixed sequence on a stable template stays efficient. The API is the right tool for a deep, permanent integration with a PDM system.
What CADABRA replaces is the automation you never wrote: the long tail of variations where scripting was never worth the maintenance.
Automation, answered.
Why not just write a macro or an API script?
Can it generate drawings for every part in an assembly at once?
Can it find the right component instead of me modelling one?
Does CADABRA integrate with PLM or PDM?
How accurate is it, and who checks the work?
How much time does it actually save?
Point it at the job you never scripted.
Bring the workflow you gave up automating and we will run it in 20 minutes.