parlay-api

Python SDK for parlay-api.com, a drop-in replacement for the-odds-api.com with 22 sources, player props, prediction markets, and WebSocket streaming.

pip install parlay-api

from parlay_api import ParlayAPI

client = ParlayAPI(api_key="YOUR_KEY")
events = client.odds("baseball_mlb", regions="us", markets="h2h,spreads,totals")

Why

The-odds-api is fine but ships flat -115/-115 placeholders for half the props you actually want, costs 5 to 6 times more at the Enterprise tier, and has no streaming. ParlayAPI ships real American odds across 22 sources, integrates Polymarket as a sharp prediction-market baseline, and runs WebSockets on Business+ tiers.

The SDK preserves TOA's endpoint shape, so migration is one config change.

Migration from the-odds-api

If you're currently using the unofficial Python client:

# before
from the_odds_api import OddsAPI
client = OddsAPI(api_key=KEY)
events = client.get_odds(sport_key="basketball_nba", ...)

# after, three-line diff
from parlay_api import ParlayAPI
client = ParlayAPI(api_key=KEY)
events = client.odds("basketball_nba", ...)

If you're calling TOA via raw HTTP, just swap the base URL:

# before
url = "https://api.the-odds-api.com/v4/sports/basketball_nba/odds"

# after
url = "https://parlay-api.com/v4/sports/basketball_nba/odds"

Same paths (/v4/sports, /v4/sports/{sport}/odds, /v4/sports/{sport}/events, /v4/sports/{sport}/scores, /v4/historical/sports/{sport}/odds), same response shapes, same query parameters (regions, markets, bookmakers, oddsFormat, dateFormat, eventIds, commenceTimeFrom, commenceTimeTo).

What's different from TOA

Beyond TOA's endpoints, ParlayAPI exposes:

Plus markets=player_* works on the standard /v4/sports/{sport}/odds endpoint, returning paired Over/Under in TOA's exact event shape.

Quickstart

from parlay_api import ParlayAPI

client = ParlayAPI(api_key="YOUR_KEY")

# List sports
print([s["key"] for s in client.sports()])

# h2h + spreads + totals across every book
events = client.odds(
    "basketball_nba",
    regions="us",
    markets="h2h,spreads,totals",
    odds_format="american",
)

# Player props (TOA-shaped)
prop_events = client.odds(
    "baseball_mlb",
    markets="player_strikeouts,player_total_bases,player_hits",
)

# Player props (flat list, ParlayAPI-extension)
rows = client.props(
    "baseball_mlb",
    markets=["player_strikeouts"],
    bookmakers=["draftkings", "pinnacle", "fanduel", "bet365"],
)
for r in rows[:10]:
    print(r["bookmaker"], r["player_name"], r["line"], r["over_price"], "/", r["under_price"])

# Devig a paired market on the client side, no network
fair_over, fair_under = ParlayAPI.devig(over_price=-110, under_price=-110)
# fair_over == 0.5, fair_under == 0.5

# Compute a +EV edge in percentage points
edge = ParlayAPI.edge(book_price=+120, fair_prob=0.55)
# +9.5pp positive expected value

WebSocket streaming (Business+)

pip install parlay-api[ws]

import asyncio, json, websockets
from parlay_api import ParlayAPI

client = ParlayAPI(api_key="YOUR_KEY")

async def run():
    async with websockets.connect(client.websocket_url("baseball_mlb")) as ws:
        while True:
            msg = json.loads(await ws.recv())
            print(msg.get("type"), len(msg.get("data") or []))

asyncio.run(run())

You'll receive an initial snapshot frame (type=odds_update with the full current state) followed by push frames whenever any book updates a price. Optional event-level filter:

await ws.send(json.dumps({
    "type": "subscribe",
    "event_id": "2026-04-29_New_York_Yankees_Boston_Red_Sox",
}))

Examples

The examples/ folder ships three runnable scripts:

• ev_scanner.py — pulls a sport's player props, devigs the sharpest book in each market as the fair-price baseline, and prints every other book's posted line that exceeds a +3pp edge threshold. This is the basic shape of a +EV scanner.
• arb_finder.py — scans h2h game lines across every book in the response for any pair where combined implied probability is under 100%. Computes the optimal stake split for a given bankroll. Risk-free profit when an arb exists.
• websocket_stream.py — subscribes to the live odds WebSocket for a sport and prints diff frames as they arrive.

Run any of them after setting PARLAY_API_KEY:

export PARLAY_API_KEY=your_key
cd examples
python ev_scanner.py baseball_mlb player_strikeouts --edge 3.0
python arb_finder.py basketball_nba --bankroll 1000
python websocket_stream.py icehockey_nhl

Devig math

The SDK includes the four standard devig methods so you don't have to reimplement them:

ParlayAPI.american_to_implied(-110)   # 0.5238
ParlayAPI.american_to_decimal(-110)   # 1.9091
ParlayAPI.decimal_to_american(2.10)   # 110
ParlayAPI.implied_to_american(0.55)   # -122

# Devig with three different methods:
ParlayAPI.devig(-110, -110, method="multiplicative")  # standard
ParlayAPI.devig(-110, -110, method="additive")        # less accurate
ParlayAPI.devig(-1500, +700, method="power")          # better for skewed

Error handling

from parlay_api import (
    ParlayAPI,
    ParlayAPIError,
    InvalidAPIKeyError,
    CreditLimitExceededError,
    RateLimitedError,
    TierGatedError,
)

client = ParlayAPI(api_key="...")
try:
    events = client.odds("basketball_nba")
except CreditLimitExceededError:
    print("hit monthly credit cap")
except RateLimitedError:
    print("slow down, retry in 1s")
except TierGatedError:
    print("upgrade required for this feature")
except InvalidAPIKeyError:
    print("API key revoked or wrong")
except ParlayAPIError as e:
    print("other error:", e)

After every successful call the SDK exposes the credit headers via client.last_quota:

client.odds("basketball_nba")
print(client.last_quota.requests_remaining)  # int

Pricing

Sign up at parlay-api.com.

License

MIT. See LICENSE.

Bugs and support

github.com/JacobiusMakes/parlay-api-python/issues or email peakpotentialmediaventures@gmail.com.