SDS for Novel Compounds — In Minutes, Not Weeks.

A proxy SDS (also called a read-across SDS) is a Safety Data Sheet generated for a novel compound by borrowing safety data from a structurally similar, well-documented compound. Rather than leaving your lab with no documentation at all, it provides a conservative, science-based estimate of hazard properties. ChemEngine uses ECHA's Read-Across Assessment Framework (RAAF) methodology and only bridges data from compounds with Tanimoto similarity ≥0.85 — the accepted threshold for high-confidence structural analogy. Every section that uses estimated data is clearly labeled.

Yes — within scope. For internal R&D and laboratory handling purposes, a proxy SDS generated via documented read-across methodology satisfies OSHA's HazCom requirement that employees have access to chemical hazard information. The key conditions: (1) it must be clearly labeled as a proxy/estimated document, (2) the methodology must be documented (ours uses ECHA RAAF, disclosed per section), and (3) it is not suitable for commercial distribution under GHS/HazCom regulations, REACH registration, or transport of materials for commercial sale. Your EHS team should review prior to use as official lab documentation. We include a mandatory 3-step acknowledgment in the generation flow that captures these limitations.

ChemEngine accepts five input types for novel compounds: (1) SMILES string — preferred, provides the best Tanimoto scoring accuracy; (2) InChI or InChIKey — converted to SMILES server-side via RDKit; (3) Partial or non-standard CAS — used as fallback text search in PubChem; (4) Compound name or internal designation (e.g., 'Compound 23B', 'NVX-301') — triggers PubChem fuzzy name search; (5) Molecular formula (e.g., C₁₄H₁₉N₃O₂) — lowest precision, generates multiple proxy candidates for you to select from. SMILES is strongly recommended for unambiguous compound identification and highest proxy match quality.