Spark starts immediately
Open a managed account, build a focused first model, and keep local file export available.
Production BBN workspace
Darkstar Online
Scalable causal learning and inference at production speed.
Build Bayesian belief networks in the browser, run exact causal queries, and move from a no-card Spark workspace to paid cloud save, sharing, support, and commercial SDK delivery.
- No card
- Spark entry
- 13 ports
- SDK archive matrix
- Protected
- Runtime grants
Paid plans add persistence
Cloud documents, autosave, larger limits, sharing, and authenticated support unlock as work grows.
Runtime access is gated
Protected WASM grants tie inference access to the account and entitlement state.
Workspace
A dense modeling surface for repeated analytical work.
Graph authoring
Build directed acyclic graphs, edit node states and probability tables, and keep layouts readable as networks grow.
Exact queries
Inspect evidence, posteriors, lift, interventions, and counterfactuals without leaving the workspace.
Data sampling
Generate sample rows, test scenarios, export synthetic data, and inspect how assumptions move through the model.
Structure and parameter learning
Planned cloud-native learning will add constraint, search-and-score, and optimization workflows to the same product surface.
Example models
Start from real network shapes.
The public model corpus gives Spark users meaningful documents to open, inspect, query, and export before committing paid cloud storage.
Huang diagnosticLeukemiaAutismVAERS
Protected runtime
Access follows the account.
Spark, paid plans, support credits, and commercial SDK licenses all resolve through the same account boundary, so product access stays explainable.
- Managed social sign-in
- Spark entry without card collection
- Clear subscription recovery path
- Signed SDK artifacts and manifests