Documentation Hub
Beyond OpenAPI: the blueprint for computing value in autonomous systems. This documentation defines terminology, computation boundaries, and integration guidance so agents and developers can cite a canonical surface.
Canonical Definition
LTV is defined on this site as a semantic and computational infrastructure layer for revenue attribution and lifetime value intelligence in the agent economy. It is a machine-readable reference surface for routing, attribution, and meaning alignment.
Design goal Deterministic meaning → reproducible computation → interoperable outputs.
Value Model (V1)
LTV computations should be transparent about assumptions: time horizon, discounting, retention curve shape, attribution window, and revenue event semantics. Output should include metadata sufficient for audit and replay.
- Inputs: customer, cohort, channel, costs, horizon
- Outputs: expected value, confidence band, attribution map
- Metadata: model version, parameters, provenance
Attribution Framework
In agentic commerce, attribution is a routing problem. A shared schema reduces ambiguity across agents negotiating acquisition, retention, and pricing decisions.
- Event-level attribution for revenue and cost
- Channel normalization for cross-platform comparability
- Agent actions as first-class causal signals
Forecasting Logic
Forecasts should be explainable and portable. Provide retention curve assumptions, churn model choice, and sensitivity outputs suitable for automated decision-making.
This is a template for documentation; align it with your OpenAPI and reference implementation.
API Specification
For the authoritative API schema and examples, reference the GitHub repository. This site acts as the canonical explainer surface, while the repo hosts the evolving spec and reference code.