Probabilistic Physics-Based LMP Forecasting

Model the Physical Grid

Distill runs optimal power flow on a full topology representation: every line, every bus, and every transformer. Unlike machine learning models that struggle to generalize to unseen or unexpected scenarios, our physics-based model adapts to the future because it's solving the actual physics problem.

Honest About Uncertainty

Distill is built on the fact that when load, fuel, weather, and outages are all uncertain, the LMP is more uncertain than any one of them. Every downstream quantity is inherently probabilistic and uncertainty is quantified.

Any Scenario in Minutes

Distill empowers you to alter the gas price curve, change the wind shape, retire a plant, add a data center, swap a transmission outage in and out, and re-run a year of hourly nodal simulation in minutes.

Integrated into your workflow

Distill is a tool you adopt without requiring months of training sessions. APIs and MCPs integrate into existing workflows and empower new ones. We will never bottleneck your work with desktop applications, antiquated workflows, or lengthy implementation cycles between signing and using.

Full-scale power flow in your browser.

Distill is the first platform to run complete, full-fidelity nodal power flow from a web browser. The compute lives in our cluster and is only one URL away.

Cloud-native, parallel by default.

Distill runs entirely on cloud infrastructure built for parallel execution. A hundred probabilistic scenarios finish in the time of a single deterministic run.

Queryable and accessible outputs.

Every simulation can be queried efficiently with SQL or imported into your tool of choice.

Load a pre-assembled topology, like

Edit any feature, like

Create, use, or sample from forecast quantities, like

Choose from forecast vendors, like

— or bring data from your own sources.

Distill is an interconnected platform. Click a node to explore. Drag to rearrange.