Turn design state into a
learnable product layer
AiEDA is iEDA.ai's intelligence layer: Library exports vectors and graphs, Models embed in flow decisions, Dataset ensures reproducibility, iPCL targets layout pre-training.
# design state → vector → model → feedback [library] features & graphs [model ] task heads in-flow [data ] open reproducible sets → agents observe the same iDB
Library
Design-to-Vector: extract cells, nets, timing, congestion, and other structured features from any intermediate state.
Model
Task heads such as AiMap/AiSTA/AiTO predict and guide before expensive steps.
Dataset
Open chip/problem/tool-level data for fair comparison.
iPCL
Layout pre-training: representation · generation · metric evaluation.
Module entry
Enter by product capability, not course catalog.
AiEDA Library
Python API: install, getting started, and interfaces.
Enter →Task model matrix
Six in-flow model cards.
Enter →Open datasets
Chip, problem, and tool data entry.
Enter →Layout pre-training
R/P/M three-head product pages.
Enter →AiEDA TCAD paper
Design-to-Vector system work.
Details →Back to platform
Unified database the intelligence layer depends on.
Platform →