Design working formulations before your first batch
Celleste charts the formulation space for lab-grown materials and forecasts which recipes can perform before you commit to a batch. Get a validated sample in months, not years.
Formulation Predictor
Composition inputs
Predicted properties
Bio-material iterations often fail for reasons visible only afterward
Materials scientists creating lab-grown materials often run 40 to 80 batch iterations before finding a working formulation. Every failed batch takes weeks and uses costly precursors. Without composition-based performance predictions, each batch becomes a high-stakes guess.
Turn composition data into candidate formulations in a workflow
Add your material data
Bring in batch records, literature values, or pilot measurements. Celleste takes in composition variables together with their measured properties.
Create candidate formulations
Celleste explores your composition space and suggests formulations expected to meet your target properties, ordered by predicted performance.
Predict before batching
For every candidate, Celleste provides predicted tensile strength, biodegradation rate, and cell adhesion profile with confidence intervals. Run only formulations worth testing.
Three ways Celleste supports materials scientists
Generate formulations
Using your target property profile and lab composition variables, Celleste creates a ranked set of candidate recipes. No guessing from a blank slate.
Predict properties
Use composition inputs alone to predict tensile strength, elongation at break, cell adhesion score, and biodegradation rate. See the performance profile before running the batch.
Track iterations
Record predicted and actual results over time. As your lab gathers real batch data, Celleste sharpens its predictions and connects prediction with experiment.
Results from our internal pilot
Across 3 bio-material development cycles in our internal pilot, average iterations to a validated sample fell from 62 to 11.
Source: internal pilot data, early-access program, 2025. Bootstrapped-stage pilot, not enterprise-scale claim.
Celleste reshaped our formulation process. We once relied on intuition, then iterated until a result worked. Now we begin with ranked candidates and predicted properties. We still run batches, but we run fewer.
Dr. Maya Ohayon, Materials Scientist
composite bio-materials lab, early-access program participant
Begin with your formulation challenge
Test Celleste with material data. The Pilot plan is free with no time limit.