Public REST API

API

Normalized current data and provenance from the Django origin.

Base URL

https://api.modelmarkets.ai/api/v1

The production hostname is documented but not deployed. Local development uses http://127.0.0.1:8000/api/v1.

Response envelope

{
  "data": { ... },
  "meta": {
    "dataset_status": "sample | live",
    "is_live": false,
    "disclaimer": "..."
  }
}

v1 endpoints

Current surface

MethodPathReturns
GET/api/v1/models/Model version directory with provider counts and price ranges.
GET/api/v1/models/{organization}/{model}/Complete model record with endpoints, prices, aliases, and provenance.
GET/api/v1/providers/Inference provider directory.
GET/api/v1/providers/{provider}/Provider endpoints, model metadata, pricing, and sources.
GET/api/v1/pricing/Current normalized price records with units and source classification.
GET/api/v1/provenance/Source registry and collection metadata.
GET/api/v1/compute/Compute venue and instrument metadata.
GET/api/v1/health/API origin health.
Authentication foundation

Django users and hashed API credentials are modeled. Public read endpoints are rate-limited; credential issuance is intentionally not activated in this delivery.

Cache contract

Responses include public cache-control with stale-while-revalidate. Redis provides shared current state; the frontend uses server rendering and incremental revalidation.