Open the lot and isolate the failing population.
Use linked scatter, histogram, wafer, and spatial views to decide whether the problem is edge-driven, test-driven, or localized to a subset.
Closed alpha
Go from STDF lot data to a reviewable yield decision in one workflow.
Find whether a problem is edge-driven, test-driven, or limit-driven, stage the proposal, review exact yield impact, and save the conclusion as a notebook artifact instead of rebuilding the same context across five tools.
Already using stdf.io? Stratum is the team workflow that picks up where quick local inspection stops.
Typical review Lot 18A7 • SKEW_CLK • Yield proposal
Spatial signal Edge-driven fail pattern on W12
Worst driver SKEW_CLK Cpk 0.01
Proposed change USL 120 ps → 135 ps
Exact impact Yield 82.4% → 89.1%
Linked analysis One context, four surfaces
Scatter, histogram, wafer, and spatial views stay in the same decision loop.
Notebook artifact Something you can defend
Save the proposal, the rationale, and the reviewed impact as one shareable artifact.
Decision path
Three steps from failing lot to defensible proposal.
The goal is not more dashboards. The goal is getting to the point where an engineer can explain what is wrong, what changed, and why the proposed limit move is justified.
Stage the limit change and review exact impact.
What-if edits are not just chart interactions. Stratum recomputes rescued dies, newly failing dies, per-test deltas, and net yield impact.
Save the conclusion as a reviewable artifact.
Keep the notebook, version lineage, share links, and controlled-document workflow instead of losing the reasoning in screenshots and chat.
What makes it different
Built for engineering judgment, not just chart display.
Stratum sits between raw lot data and the moment a spec, limit, or failure explanation needs to be defended.
Persistent lot history
Lots, snapshots, normalized test metadata, and repeatable server-side analysis instead of one-off desktop exports.
Linked analysis loop
Scatter, histogram, wafer, and spatial views share context so the investigation moves forward instead of resetting on each surface.
Exact what-if review
Limit proposals are reviewed with real overlap handling and exact yield deltas, not hand-waved percentages.
Notebook artifact
The conclusion is saved with version lineage, sharing, and controlled-document workflow when a decision needs to be reviewed or frozen.
Who it is for
Mixed-signal and analog teams that need faster root-cause loops.
The alpha is deliberately narrow. It is optimized for STDF-centric lot analysis, engineering review, and practical collaboration, not for trying to be a giant operations suite on day one.
stdf.io
Fast local analysis
- Browser-only
- Single-user
- No saved server-side history
- Best for quick inspection and free exploration
Stratum
Persistent team workflow
- Server-side lots and snapshots
- Notebook-backed analysis and sharing
- Exact what-if review
- Access control and governed artifacts
- Path to SQL and Jupyter-style notebooks on the same platform
Platform direction
Yield analysis first. Broader notebook workflows next.
The current surface is deliberately focused on STDF lot analysis. The same document lifecycle is meant to extend into SQL and Jupyter-style notebooks later, but that is downstream of getting the core yield-decision workflow right.
Get oriented
Use the docs first. They are part of the product surface now.
If you are evaluating the alpha, start with the quickstart and analyze workflow. If you want access or a walkthrough, use the alpha request link.