Subspace

Subspace is a platform for executing business logic models at scale without managing infrastructure. It describes the models as declarative JSON and supports named logic in a business registry, so applications can call a model by name rather than resending a full specification each time.

The product emphasizes three points: declarative model definition, parallel execution of thousands of scenarios in a few milliseconds, and reproducibility, where the same model and the same seed produce the same results. The documentation also says that SP Models are a JSON format for data projection, and that models can be deterministic or random. Random variables may use the uniform, normal, lognormal, or exponential distributions. Example uses named in the documentation include actuarial reserve projections, finance risk analysis and stress testing, technical modeling, performance simulations, optimization, predictive projections, statistical modeling, capacity planning, supply-chain budgets, and similar projection or simulation tasks.

Access is provided through a REST API and SDKs for Python and TypeScript/JavaScript. The documentation shows API key authentication, uses the X-API-Key header for REST calls, and notes that the Developer Console is the entry point for business logic, API keys, usage, audit, and billing. It also mentions integrations with n8n and MCP, including use from Claude Desktop and Cursor via the MCP protocol. A “persist_mode” setting is available when creating an API key: minimal returns only the result while logging quota and usage, and full stores runs and replay data in Subspace.

Subspace is aimed at actuaries, engineers, analysts, and people working in finance or risk management. The page also refers to a developer workflow, with quick-start instructions for obtaining an API key, installing a SDK, and making a first call.