Real-Time Prediction Market Feed

Polling vs. streaming

Prediction market APIs are generally poll-based: you request current data on a schedule. The Meridian Edge API is optimized for frequent polling — responses are cached server-side and return in under 200ms for most requests.

Pro-tier users get access to webhook delivery, which pushes updates when events change by more than a configurable threshold — eliminating the need to poll when nothing has changed.

Choosing your polling interval

• Sports in-game (live): 60–120 seconds. Markets can shift several percent points in minutes when a game changes.
• Pre-game (same day): 5–10 minutes. Injury reports and lineup news drive meaningful moves.
• Pre-game (future): 30–60 minutes. Market-moving news is less frequent.
• Macro / political events: 1–4 hours. These markets move slower unless breaking news occurs.

Starter tier polling math

The starter tier includes 100 calls/day. With 4 sports and polling every 15 minutes for 6 hours of active game windows, that's approximately 4 × (6 × 4) = 96 calls — just within the starter limit.

Starter (1,000 calls/day) supports continuous polling every 5 minutes across all sports for a full 24-hour day.

Efficient polling pattern

import requests, time

API_KEY = "me_live_your_key"
HDRS    = {"X-API-Key": API_KEY}

# Only re-fetch if data is stale (use ETag or last-modified)
last_etag = None

def fetch_if_changed(sport="NBA"):
    global last_etag
    headers = dict(HDRS)
    if last_etag:
        headers["If-None-Match"] = last_etag
    r = requests.get("https://meridianedge.io/api/v1/consensus",
                     headers=headers, params={"sport": sport})
    if r.status_code == 304:  # not modified
        return None
    last_etag = r.headers.get("ETag")
    return r.json().get("events", [])