The toolkit

Seven tools. One workflow.

From reaction definition through pilot-scale safety review and cost-per-kg. Each tool feeds the next — no re-keying, no copy-paste between spreadsheets, no version-confusion.

Phase 01
Plan
Define your chemistry. Optimize the conditions before you commit a single drum.
⚗
Reactions
Foundation
Stoichiometric tables with PubChem-backed compound lookup. Auto-calculate masses, volumes, and equivalents — then reuse the recipe everywhere else in the suite. Save templates for chemistry your team runs often.
  • Real-time mass & volume calc
  • Reusable templates & teamspace
  • Feeds every downstream tool
∿
Experiment Optimizer
Bayesian · self-tuning
Stop running 30-experiment factorial campaigns. Define your variables, objectives, and constraints, paste in what you have, and the self-tuning Gaussian-Process optimizer recommends the next batch that closes the most uncertainty per run. On a 1,000-candidate benchmark it reaches the optimum in under 5% of experiments — typically saving $90K+ per screen at $200/experiment.
  • Self-tuning model — multi-restart MLE + per-dimension lengthscales
  • Multi-objective (yield + time + selectivity), constraints, replicate handling
  • Outlier detection, CV diagnostic, target-reached stopping signals
  • Team workspaces with role-based permissions
Phase 01 spotlight · Experiment Optimizer
Under 5% of experiments to optimum.

On a 1,000-candidate design (5 catalysts × 5 solvents × 10 temps × 4 pressures), the self-tuning Gaussian Process reaches 99% of the true optimum in just 3.6–4.1% of experiments — about 40 picks. A first-generation Bayesian optimizer plateaus at ~83% yield and never reaches the optimum, even after 500 picks.

vs. brute force
~$192K
saved per 1,000-cell screen at $200/exp
Yield only — single objective 40 50 60 70 80 90 0 25 50 75 100 true optimum 91.4 optimum @ exp 41 Yield max + time min — multi-objective 0 25 50 75 100 true optimum 91.4 optimum @ exp 36 Experiments run Best yield observed ScaleChem v13 First-generation BO
What's under the hood: multi-restart MLE with per-dimension lengthscales (ARD) · Augmented Expected Improvement · Kriging Believer batch · adaptive diversity weight · constraint support · replicate handling · outlier detection · cross-validation diagnostic · multi-objective with goal toggle · team workspaces.
Phase 02
Predict
Will it work at 100×? Quantify the thermal, mixing, and yield risk before the pilot run.
🜂
Heat Transfer Modeling
Thermal safety
Quantify thermal risk before scale-up. Combines reaction enthalpy, kinetics, jacket heat-removal capacity, and Stoessel runaway classification. Six pillars including operator SOP charts and reality-check calibration from observed batch data.
  • q_peak vs cooling capacity at every scale
  • MTSR & Stoessel class with confidence bands
  • Operator SOP: addition rate vs bulk T
  • Calibrate UA from your observed runs
∝
Dimensionless Scaling
Mixing & transport
Reynolds, Power number, mixing time, Damköhler, and tip speed — automatically computed across lab, pilot, and plant scales. Verify regime similarity before you commit capital. Custom Cp library shared across the suite.
  • Pick a scaling criterion, see consequences
  • Three scales side-by-side, instantly
  • Custom impellers & jacket fluids
◉
Yield Prediction at Scale
Loss-mechanism modeling
Predict yield drop at scale by mechanism — mixing, thermal, holdup, workup, mass-transfer, stoichiometry. Honest about uncertainty: confidence-tiered ±5–15% bands calibrated against published case studies. Auto-links to your dimensionless and heat-transfer analyses so the predictions reflect your actual setup, not generic correlations.
  • Per-mechanism breakdown of expected loss
  • Sensitivity to addition time, agitator, cooling
  • Confidence bands, not point estimates
  • Auto-pulls your HT & dim numbers — no re-keying
Phase 03
Build
Pick the right reactor and the right cooling. Justified by your upstream numbers.
⚙
Equipment Recommender
Reactor & agitator picker
Synthesizes your heat-transfer, mixing, and yield analyses into concrete recommendations at every target scale: reactor type, agitator, operating mode, cooling strategy, even chiller wattage. With justifications referencing your upstream values, not generic rules-of-thumb.
  • Glass / steel / Hastelloy / continuous flow
  • Chiller spec: capacity, setpoint, water flow
  • Cross-scale transition warnings
Phase 04
Safety
Determine which regulations apply, then build the PHA your team will actually use.
⚠
Process Safety — Standards & PHA
Applicability engine + HAZOP-style PHA
Auto-evaluates your reagent inventory against the canonical published threshold lists encoded in regulations.js — federal and state — then walks you through a HAZOP-style PHA on a calibrated 5×5 risk matrix. ~180-entry embedded chemical library auto-fills hazard data and threshold quantities. Treats the verdict as a heuristic input to your formal applicability determination, not a replacement for it.
  • Federal lists encoded: OSHA 1910.119 App A (133 entries), EPA 40 CFR §68.130 Table 1 RMP toxic (77), Table 3 RMP flammable (62), 10,000 lb aggregate flammable trigger
  • State-specific encoded: CalARP 19 CCR §5130.6 Table 3 (full 272 entries), NV CAPP NAC 459.9533 (full 213), NJ TCPA N.J.A.C. 7:31-6.3 (8 NJ overrides), DE EHS (Acute Toxicity Concentration methodology), Contra Costa County ISO
  • Other: CFATS 6 CFR Part 27 App A (partial subset). Aqueous-vs-anhydrous handling per OSHA enforcement.
  • Exemptions encoded: R&D exemption (PSM 1910.119(a)(2)(iii) + RMP §68.115(b)(5)), atmospheric tank exclusion, hydrocarbon fuel exclusion
  • HAZOP-style PHA: 30-60 scenarios per process, heuristic generator, calibrated 5×5 risk matrix. Smart-suggestions flags DMF + strong-base, pyrophorics, peroxide formers, etc.
  • CSV worksheet + printable HTML report for the team meeting. AI-assist (Pro) for novel chemistry.
Phase 05
Economics
Once it scales, works, and is safe — what does it actually cost?
$
Techno-Economic Analysis
Cost per kg
Estimate cost per kilogram at production scale. Multi-stage synthesis, recycle modeling, equipment sizing, and tornado sensitivity analysis — in one tool. Get your "what does this cost at 10,000 kg/year?" answer before the meeting, not after.
  • Multi-stage synthesis & recycle accounting
  • Capex / opex / labor / utilities
  • Tornado plots for sensitivity

See it on a real reaction.

A worked example: SN2 alkylation of diethyl malonate, walked through every tool from 1L bench to 1500 kg/year — with real numbers and real recommendations.

Open the demo → Our products