Two workflows in one tool: APRS portfolio pre-fill, and a formulation-hazard extraction prototype (Stefan use case 2).
Shared input for both workflows: APRS portfolio pre-fill and the formulation-hazard prototype. Upload once, then switch tabs below.
After extraction, each ingredient is matched against the shared template schema here — resolve Unmapped values before writing any workbook.
When you pick an option for an Unmapped endpoint, that choice is saved (schema + endpoint + extracted text). The next time the same text appears for that endpoint, it is applied automatically as remembered — you will not need to pick again. Export/Import carries the memory between browsers.
General principles you approve in chat (e.g. Section 2 silence → not classified) are saved here and sent with the next extraction. Export writes Markdown matching knowledge/reviewer-memory.md.
Run at least one ingredient above to extract and schema-match endpoints here.
Writes the already-resolved classifications into one shared workbook (the starting point every country copy is cloned from). Does not re-run schema matching — resolve everything in Section 2 first.
Once the shared workbook is generated, make N identical copies — one per country/template group (~13 focus countries share ~8 templates today). Each copy starts identical; label it, then download it. Country-specific changes (dropping ingredients, adjusting scores) happen later, by hand, in Excel.
Generate the shared workbook above to enable this.
Same MSDS files as section 1. Enter each component’s % w/w, extract six CLP endpoints (classified / not classified / no data), then download a pre-filled ClassCalc AcuTox workbook (INPUT sheet). ATEmix / GCL / SCL stay as Excel formulas — open the file in Excel to see PROJECT results.