Structure Editor
Draw or paste a molecule once and reuse it across ChemistryAtlas tools.
Browse ChemistryAtlas generative molecular design tools for molecule generation, property-constrained design, virtual libraries, docking and active-learning screening, ChemProp-style prediction, MolPAL workflows, and model connectors.
These apps are the pro tier for ideation, library growth, multi-objective scoring, and AI-assisted molecule optimization.
These apps are the pro tier for ideation, library growth, multi-objective scoring, and AI-assisted molecule optimization.
The Generative Molecular Design Pro hub focuses on advanced AI workflows: target-conditioned molecule generation, virtual library enumeration, active-learning screening, docking connectors, ChemProp-style property models, and model-backed design loops.
Draw or paste a molecule once and reuse it across ChemistryAtlas tools.
Run RDKit Morgan-fingerprint similarity and SMILES/SMARTS substructure matching over a pasted or uploaded library.
Detect common retrosynthetic disconnections such as esters, amides, aryl halides, and boronic-acid coupling handles.
Apply RDKit reaction SMARTS templates for amide, ester, and hydrogenation product prediction.
Extract Murcko scaffolds and scaffold atom counts from names or SMILES with RDKit.
Enumerate and rank substituent designs against MW, TPSA, and target-logP constraints.
Build pairwise SAR delta tables with Murcko scaffold checks, Tanimoto similarity, and property changes.
Select diverse compounds by greedy max-min distance over RDKit Morgan fingerprints.
Enumerate R-group or reaction-SMARTS virtual libraries, calculate product descriptors, and export SDF/CSV.
Screen molecular precursors for linker, electrolyte, chromophore, and organic-electronics motifs.
Generate candidates toward logP, MW, TPSA, solubility, pKa, ADMET, or custom property goals.
Propose scaffold replacements that preserve property profile while changing core structure.
Enumerate and rank substituent replacements using RDKit descriptors, synthesis filters, and optional ChemProp predictions.
Optimize candidate molecules across potency proxy, solubility, alerts, diversity, novelty, and synthesis feasibility.
Generate molecules and require route/template feasibility before ranking final candidates.
Suggest medicinal-chemistry transformations and property-direction changes from matched-pair rules.
Upload measured data and candidates, train a local model, and recommend next molecules to make/test.
Build focused virtual libraries from cores, R-groups, reaction handles, and property constraints.
Train property models using ChemProp sidecar when available or local RDKit Random Forest fallback.
Run batch molecular property prediction with saved ChemProp models or local descriptor fallback.
Estimate uncertainty and applicability domain from ensemble/model spread or fingerprint distance fallback.
Predict MS/MS fragments using ms-pred/ICEBERG sidecar when available or deterministic exact-mass losses locally.
Score experimental vs predicted/library spectra with cosine similarity and matched-peak diagnostics.
Rank candidate structures against formula, exact mass, spectrum hits, fragments, and property filters.
Combine formula constraints and MS/MS similarity to produce structure-elucidation candidate reports.
Prepare a docking workflow and call PyScreener sidecar when configured; otherwise emit runnable docking manifests.
Plan or run MolPAL-style active-learning virtual screening from docking scores and candidate libraries.
Generate a protein/ligand preparation checklist for docking, protonation, waters, cofactors, grid box, and controls.
Rank docking results with score, ligand efficiency, strain, alerts, diversity, and interaction notes.
Convert docking hits into next design ideas with R-group/scaffold suggestions and active-learning priorities.