Who does the modeling
Expert-driven GUI sessions, scripts and recipes
An AI engineer runs the method; your experts approve each stage
Netlists & test setups
Written by hand — per device, bias, temperature and simulator
Generated automatically from the measurement set and model registry
Model coverage
Turnkey packages per model family
Any standard compact model — one engine
Your own methodology
Encoded by experts in proprietary scripts
Described once in plain language, then executed every run
PDK library updates
Re-extract, then hand-merge into the library
Only the model that changed is updated, in place
Reports
Plots and decks assembled by hand
Sign-off reports with every run, in your template
Hand-offs
Separate teams and toolsets, files passed between them
One data chain from measured silicon to released PDK
Where it runs
Workstation seats operated interactively
On-premises or cloud, hosted or fully local language model
Interface
Desktop GUIs and scripting languages
Plain-language chat, web cockpit, REST API — zero scripting