Developer experience

Call it. Nothing here is a mock.

4 of 7 endpoints exist in this deployment and answer for real, including the refusal an integrator has to handle. The rest are marked planned or blocked, each with the dependency it is waiting on, because a technical buyer always calls the thing you show them.

Explorer

Live requests against this deployment, over the composite demonstration portfolio.

Every property in the account, with coverage and score-readiness.

/api/enterprise/properties

Response fields

properties[].id: string; Stable property identifier.
properties[].coverage: number; Share of dimensions carrying a counted reading.
properties[].score_ready: boolean; False when the engine refuses for any guest.
properties[].confidence: number; 0–100, derived from evidence grade.
versions: object; The version tuple every number was produced under.

The full surface

Marked honestly. Available means callable today; blocked names the dependency rather than a date.

ParaVida integration endpoints
Endpoint	Purpose	Status
`GET /api/enterprise/properties`	Every property in the account, with coverage and score-readiness.	Available
`GET /api/enterprise/properties/{id}`	One property's full environmental profile: every reading with its origin, grade, source, confidence and verification state, plus its Environmental DNA.	Available
`GET /api/enterprise/compatibility`	Score one property against one guest profile. Returns the engine's own refusal when coverage is short — a refusal is a valid answer an integration must render, not an error.	Available
`GET /api/enterprise/graph`	The environmental graph around a property: its neighbourhood, its destination, and the real venues that make up the environment.	Available
`GET /api/enterprise/explanation`	Guest-facing prose for a property score — strengths, the honest watch-out, and the confidence sentence.Depends onThe explanation engine is destination-shaped at BRIEFING 1. Generalising it to an arbitrary environmental entity is real work, and generating prose for a hotel without it would mean writing sentences the engine did not produce.	Requires platform work
`POST /api/enterprise/verification`	Confirm, correct or reject a reading, and re-run the affected assessment.Depends onPartner-scoped write path with an audit trail, reviewer identity, and a re-assessment pipeline triggered on evidence change.	Requires platform work
`POST /api/enterprise/webhooks`	Notify a partner system when evidence changes and a score moves.	Planned

Available: Live in this build. The explorer on this page calls it for real.
Planned: Designed against interfaces that already exist internally. Not exposed yet.
Requires platform work: Cannot exist until a named platform dependency is built. Not a scheduling gap.

The refusal contract

The part most integrations get wrong.

A refusal is a 200, not an error. It is an answer an integration must render.

When weighted coverage falls below the floor, the compatibility endpoint returns HTTP 200 with scored: false, a machine-readable reason, an explanation in plain language, and the coverage figure so the surface can show how far short it fell. It is not a 404 and not a 422, because nothing went wrong: ParaVida declined to estimate a value it has not verified, and that is the product behaving correctly.

A partner surface that assumed a number would break the first time it met an unevidenced property, which in a real portfolio is week one. Try it in the explorer above.

This API is not production-ready, and is not presented as such

The compatibility endpoint is unauthenticated and read-only against synthetic inventory. It is not a production integration surface. API key issuance, scoping, rotation and per-tenant rate limiting. It is a demonstration surface over synthetic inventory. Nothing about the shape of the interface changes when authentication exists — but nothing should be integrated against it until it does.

See how an integration is built See the readiness assessment

Produced underscoring 1 · weights 4 · ontology 1 · climate model 1 · engagement 1 · min coverage 0.5

Endpoint

Purpose

Status

GET /api/enterprise/properties

Every property in the account, with coverage and score-readiness.

Available

GET /api/enterprise/properties/{id}

One property's full environmental profile: every reading with its origin, grade, source, confidence and verification state, plus its Environmental DNA.

Available

GET /api/enterprise/compatibility

Score one property against one guest profile. Returns the engine's own refusal when coverage is short — a refusal is a valid answer an integration must render, not an error.

Available

GET /api/enterprise/graph

The environmental graph around a property: its neighbourhood, its destination, and the real venues that make up the environment.

Available

GET /api/enterprise/explanation

Guest-facing prose for a property score — strengths, the honest watch-out, and the confidence sentence.Depends onThe explanation engine is destination-shaped at BRIEFING 1. Generalising it to an arbitrary environmental entity is real work, and generating prose for a hotel without it would mean writing sentences the engine did not produce.

Requires platform work

POST /api/enterprise/verification

Confirm, correct or reject a reading, and re-run the affected assessment.Depends onPartner-scoped write path with an audit trail, reviewer identity, and a re-assessment pipeline triggered on evidence change.

Requires platform work

POST /api/enterprise/webhooks

Notify a partner system when evidence changes and a score moves.

Planned