API (v1)

GET https://nextcomp.ai/api/v1/index.json — top-level rollup: every model with KPIs, comparative spreads, sold/movers globals.

GET https://nextcomp.ai/api/v1/{slug}.json — per-model bundle: market_state, pricing_compass, active listings with buy signal + cut probability, mileage tiers, color rollup, spec premium, recent movers, BaT summary.

GET https://nextcomp.ai/api/v1/dealers.json — 312-dealer leaderboard with archetypes + trust score + flags.

GET https://nextcomp.ai/api/v1/bat.json — BaT cross-reference: resurfaced VINs, repeat-VIN history, seller scorecard, bid log analytics.

{
  "state": "Stable",
  "liquidity_score": 44,
  "median_ask": 319998,
  "confidence_adjusted_price": 319900,
  "bat": {
    "bat_model_query": "ferrari-296",
    "median_sold_6mo": 297000,
    "rnm_rate": 0.49
  }
}

Update cadence — bundles refresh nightly via the nextComp launchd job. generated_at ISO-8601 timestamp at the root of each bundle.

Schema versioning — this is v1. Breaking changes will publish under /api/v2/. Additive fields will land in v1 without version bump.

Caching — Cloudflare edge cache, ~1h TTL. If you need real-time, you're looking at the wrong endpoint anyway (this is a daily-batch dataset).

VIN data caveat — VINs surface in relisted_vins and bat.resurfaced only. These are scraped from public BaT auction pages where VINs are always public.

Read-only access is free. If you're building a product on top of this data, ping us (GitHub issues) so we know about you and can give a heads-up before schema changes.

Respect the edge cache — there's no upside to bursting requests. The data only changes once a day.