A play-by-play feed for your analytics stack

Coverage

Play-by-play: NFL, college football, NBA, MLB, NHL and 18 soccer competitions (soccer plays are match events: passes, shots, fouls, cards). Win probability series: NFL, college football and MLB. Box scores with player stats: NFL, college football, NBA and MLB. Verified on the dates shown on the endpoint pages.

Stream plays as they happen

Subscribe once per league. Every event_play message carries the game id, so you can append to a plays table keyed by game and play id. Opt in to event_play_corrected to replace plays the source later revises.

// Server-side only: your API key never ships to browsers or apps.
const res = await fetch('https://www.realtimesportsapi.com/api/websocket/auth', {
  method: 'POST',
  headers: { Authorization: `Bearer ${process.env.RSA_API_KEY}` }
});
const { data } = await res.json();          // data.url, data.token (valid 1 hour)
const ws = new WebSocket(`${data.url}?token=${data.token}`);
ws.onopen = () => {
  for (const event of ['event_play', 'event_play_corrected', 'event_final'])
    ws.send(JSON.stringify({ type: 'subscribe', event, filters: { sport: 'football', league: 'nfl' } }));
};
ws.onmessage = async (e) => {
  const msg = JSON.parse(e.data);
  if (msg.type === 'event_play' || msg.type === 'event_play_corrected') await upsertPlay(msg.data); // your warehouse
  if (msg.type === 'event_final') await reconcile(msg.data.eventId);
};

Reconcile and backfill over REST

After a final, one call returns every play and one returns the box score. For a season backfill, list games with the schedule endpoint and pull each completed game the same way.

import os, requests

BASE = "https://www.realtimesportsapi.com/api/v1/sports/football/leagues/nfl"
H = {"Authorization": f"Bearer {os.environ['RSA_API_KEY']}"}

def game(event_id):
    plays = requests.get(f"{BASE}/events/{event_id}/plays", headers=H, timeout=20).json()["data"]
    box = requests.get(f"{BASE}/events/{event_id}/boxscore", headers=H, timeout=20).json()["data"]
    wp = requests.get(f"{BASE}/events/{event_id}/winprobability", headers=H, timeout=20).json()["data"]
    return plays, box, wp

Rate and volume

Growth allows 10 requests per second and Pro 20, so a backfill of a few hundred games takes minutes. Responses include X-RateLimit-* headers; over a limit you get HTTP 429 with Retry-After.

Endpoints used

• POST /websocket/auth
• GET /sports/{sport}/leagues/{league}/events/{eventId}/plays
• GET /sports/{sport}/leagues/{league}/events/{eventId}/boxscore
• GET /sports/{sport}/leagues/{league}/events/{eventId}/winprobability
• GET /sports/{sport}/leagues/{league}/seasons/{season}/schedule

Full reference: docs · OpenAPI · what each endpoint returns, by league

Plans

Every successful REST call counts as one call. Over the limit you get HTTP 429 with a Retry-After header. No SLA. Full pricing.

FAQ

Is there tracking or advanced (expected points) data?

No. The feed is play-by-play, box scores, win probability for NFL, college football and MLB, schedules, rosters and injuries. Build advanced metrics on top of the plays.

How far back does the data go?

Completed games in the current data window are available over REST; there is no bulk historical archive download. Stream the current season to build your own history.

Can I export to CSV?

The API returns JSON. Schedule pages on this site offer JSON and CSV downloads for fixtures; for plays, write the JSON to your own tables.

Complete builds to start from

• Analyze a season of games with Python and pandas
• Get live scores into Google Sheets

Other use cases

• Score alerts for your sports app, without polling
• Live scoreboards and game pages for news and media sites
• Player stats and box scores for fantasy platforms
• Live scores on the screens in your venue
• Stream live scores to your backend over WebSocket
• Sports data for a startup, from prototype to scale