Developers

The Inferenstack API

The same numbers as the site, at machine scale — every response carries as_of, provenance, and freshness. Read-only, versioned at /api/v1.

Endpoints

Method	Path	Summary
GET	/api/v1/health	Liveness probe.
GET	/api/v1/gpus	Canonical GPU dimension + specs.
GET	/api/v1/gpus/{gpu_id}	One GPU + its latest rental-board summary (301 from an alias).
GET	/api/v1/models	Canonical model dimension.
GET	/api/v1/models/{model_id}	One model + latest token price per provider.
GET	/api/v1/providers	Provider dimension + intelligence (models served, cheapness rank).
GET	/api/v1/rental-prices/latest	Latest rental board (gpu × provider × tier × isolation × price_basis).
GET	/api/v1/rental-prices/history	Hourly/daily rental aggregate for one GPU (tier-windowed).
GET	/api/v1/rental-prices/changes	Recent rental price changes (movers).
GET	/api/v1/liquidity/latest	Per-marketplace GPU liquidity depth.
GET	/api/v1/token-prices/latest	Latest token prices per model × provider.
GET	/api/v1/token-prices/history	Daily median token prices for one model (tier-windowed).
GET	/api/v1/token-prices/spread	Per-model provider count + min/median/max in/out.
GET	/api/v1/benchmarks	Community perf per model × gpu × backend × quant.
GET	/api/v1/indices	Index catalogue + methodology version + latest level.
GET	/api/v1/indices/{index_id}/levels	Weekly index level series (anon: headline only).
GET	/api/v1/fx	Exchange rates.
GET	/api/v1/provenance/{raw_id}	Archive index row for a raw document (metadata only; anon 20/day).
GET	/api/v1/coverage	Per-source coverage stats.
GET	/api/v1/freshness	Per-source cadence + staleness.
POST	/api/v1/access-requests	Request API access (POST).
GET	/api/v1/openapi.json	This document.

Keys authenticate with Authorization: Bearer ifs_live_…. Key management lives in the console (coming soon).

Tell us what you're building.