Use a structured adaptive result

Update a documented adaptive policy from ordered observations and consume its current result directly in application code.

Outcome

Your application sends structured observations and receives the current adaptive-policy result without requiring generated language.

The structured object is the delivery contract, not the adaptive mechanism. The capability is the policy state that changes across ordered observations and remains available to later reads.

Workflow

declare domain -> reset known state -> send ordered observations -> read prediction -> validate or act -> continue

Gateways

Step	Route	Result
Create the adaptive domain	POST /api/v1/domains	Domain and adaptive template registered
Reset the policy state	POST /api/v1/domains/{domain_id}/adapt/reset	Known starting state
Add one observation	POST /api/v1/domains/{domain_id}/adapt/events	Updated prediction.<output_key>
Read without adding an observation	POST /api/v1/domains/{domain_id}/adapt/predict	Current prediction.<output_key>

Input mapping

When the application’s input fields differ from the domain’s declared observation fields, an optional mapping can translate them. If the fields already match, the application can send the observation directly. Adapt-1 updates the policy state and returns the contract-defined structured prediction; another optional mapping can select or rename output fields for the consumer’s schema.

application object
  -> optional field mapping
  -> Adapt-1 adaptive event
  -> prediction.<output_key>
  -> application schema

Any mapping should translate formats, not contain the task solution. The reasoning and adaptation remain in Adapt-1.

Validation

Assert on the object under prediction.<output_key>. Check schema validity, the required business condition, and whether reset followed by the same ordered observations reproduces the expected result.

Generated language is not required. Use an explanation route to inspect structured audit data.

Route contracts

• POST /api/v1/domains/{domain_id}/adapt/events updates policy state and returns a structured prediction.
• POST /api/v1/domains/{domain_id}/adapt/predict reads the current prediction without adding an observation.
• POST /api/v1/domains/{domain_id}/events records a normal domain event.
• POST /api/v1/domains/{domain_id}/query and POST /api/v1/domains/{domain_id}/explain return domain reasoning and audit fields.
• POST /api/v1/domains/{domain_id}/feedback reports an outcome.

The Adaptive structured predictions guide explains the exposed interval-policy template contract without tying the architecture to one application domain.

Structured transition learning uses a different Domain contract. This contract predicts an observable consequence from configured inputs and an intervention. Domain query and explanation routes return support and abstention. Read Design Domains for transition learning before you select event boundaries, evidence geometry, grouping, targets, or support controls.