Calibrated MLB probabilities, piped straight into your stack.
Three REST endpoints for the market model that powers posterior.pro. Bearer-token auth, 100 req/min per key, JSON shape stable across the v1 lifecycle. A public MLB model API with a live calibration curve — the receipt for every probability you call.
per month · base subscription
per month · this add-on
per month · API tier
Pro subscription required. Create the account and activate Pro first; then return here or use the account page to add API access.
/api/v1/picks/todayTonight's STRONG/LEAN slate + v0 NRFI + pitcher-K predictions, with signal breakdown.
/api/v1/accuracyLive Brier score, win-rate by tier, model EV at close. Same numbers the public dashboard renders.
/api/v1/calibrationCalibration buckets — predicted-vs-actual rates per probability bin. Plot it however you want.
OpenAPI 3.1 spec at /api/v1/openapi.json. More endpoints will be documented here only after they are live in the versioned API.
Issue a key at /mlb/account, then pipe the secret into a Bearer header. The same request, three idioms:
$ curl -H "Authorization: Bearer pk_live_..." \ https://posterior.pro/api/v1/picks/todayconst res = await fetch(
"https://posterior.pro/api/v1/picks/today",
{ headers: { Authorization: `Bearer ${process.env.POSTERIOR_KEY}` } }
);
const { picks, ext_markets } = await res.json();import os, httpx
r = httpx.get(
"https://posterior.pro/api/v1/picks/today",
headers={"Authorization": f"Bearer {os.environ['POSTERIOR_KEY']}"},
)
slate = r.json()Response shape (truncated):
{
"api_version": "v1",
"generated_at": "2026-05-26T17:42:11.103Z",
"game_date_et": "2026-05-26",
"picks": [
{
"prediction_id": "pred_01HXJ6V…",
"batter_name": "Aaron Judge",
"team_name": "New York Yankees",
"market": "1plus_hits",
"p_predicted": 0.7124,
"sharp_consensus_prob": 0.6680,
"sharp_book_american_odds": -210,
"tier": "STRONG",
"signals": { "park_adj": 1.04, "pitcher_adj": 0.96 }
}
],
"ext_markets": { "nrfi": [ ... ], "strikeouts": [ ... ] }
}Wire Posterior into an agent, Discord bot, notebook, or internal dashboard with ordinary HTTPS fetches. Your tool asks what Posterior says about tonight's slate; the API returns model probabilities with the calibration receipt attached.
The same model that runs the public dashboard. Same calibration scaler, same sharp-consensus de-vig, same conservative risk reference sizing. No hidden entries, no premium-only model layer, no certainty framing behind the API that isn't visible on the site.
The API tier is the same data in machine form, for builders who want to wire it into their stack. The calibration receipt at /mlb/accuracy is the load-bearing claim on both sides.
Optional changelog list — one email per major endpoint release. Not required to subscribe; the existing API works today without it.