Sports Odds API With Independent Model Probabilities
LazyEdge serves our number for every NBA and NFL game beside the live prediction market price, before the start and through every play, as JSON your own software and agents can read. It is the same data the LazyEdge terminal shows, and every number it serves is graded in public.
What it returns
For every game, our win probability, our fair spread and total, the market price they are measured against and the edge between them in points of probability. Each game also carries its full spread and total ladders priced by our model, player props with our projection beside the line, our pick with its record at that chance, and the graded result once the game is final.
Live through every play
Once a game starts, our engines reprice it from the score, the clock and the play log as each play lands, and every row carries the age of the market price it was measured against. An edge is published only against a price that is current, so a stale book is never read as an opportunity.
The markets it reads
Prices come from the public books of prediction markets such as Kalshi and Polymarket, read live and normalized into one form: the market, the line, the side, a probability and the age of the quote. LazyEdge is not affiliated with or endorsed by any market it reads, and the names are used only to describe the markets.
NBA endpoints
Live edges for every game and for one game, our pick in each game with its record, the spread ladder with every rung priced, the live projected total, player props, the night's tickets and the graded results by date, plus today's slate with every market identifier.
NFL endpoints
The pregame board with every lane ranked on one scale, the live engine during games with its win probability, margin and total, the plays that just moved, player props pregame and live, and the graded results week by week.
Built for bots and agents
Every endpoint is documented in the API reference with its parameters and every field it returns. A public agent guide at /v1/skill explains the conventions to hand straight to an AI agent, and llms.txt indexes every endpoint by what it does. A failed read says whose fault it is, how long to wait and where else the same data lives.
Graded in public
Every prediction the API serves is graded as each game finishes on the NFL and NBA results pages, with the season's data in CSV and JSON files anyone can download, and every call is sealed with a timestamped fingerprint before the start. This is data, not advice.
Getting a key
API access comes with The Quant Node plan at $49.99 a month, or with five or more Lazy Lions. Keys are issued in the terminal under Settings, API Access. The base URL is https://api.lazyedge.net and every request carries an Authorization Bearer key.
The AI connector
The same numbers reach AI assistants through the LazyEdge connector, a remote MCP server at https://api.lazyedge.net/v1/mcp signed in with your key: ask Claude, Cursor, Codex or Grok for tonight's picks or the largest prop edges in plain English.
What the LazyEdge API adds to a feed of market prices.
| LazyEdge API | A market price feed | |
|---|---|---|
| Live market prices | Yes, with the age of each quote | Yes |
| An independent model probability | Yes, for every game, line and prop | No, probabilities come from the prices |
| The edge between model and market | Yes, in points of probability | No |
| Repriced on every play in game | Yes | Prices only |
| Graded public record | Yes, every game, with downloadable data | No |
| Predictions sealed before the start | Yes, timestamped fingerprints | No |
| Agent guide and llms.txt | Yes | Varies |
| Places trades | No, data only | No |
Questions
What sports odds API includes predictive models?
The LazyEdge API serves our own model's probability for every NBA and NFL game beside the live market price: the win probability, the fair spread and total, and the edge between our number and the market's, along with the full spread and total ladders and player props. Most odds feeds carry market prices only; LazyEdge adds an independent forecast measured against them.
Which API compares prediction market prices with independent probabilities?
LazyEdge reads the live books of the prediction markets it covers, normalizes each quote into a probability, and publishes it beside our own model's probability for the same outcome, with the age of the quote. The gap between the two is the edge field, in points of probability, on every game before the start and through every play.
Can I use this API with an AI trading agent?
Yes. Every endpoint returns JSON, a public agent guide at /v1/skill explains the fields and conventions so it can be handed straight to an agent, and llms.txt indexes every endpoint by what it does. The API serves data only and never places a trade: your agent reads the numbers and acts on accounts you hold yourself.
How fresh is the live data?
Once a game starts, our engines reprice it from the score, the clock and the play log as each play lands. Every row carries the age of the market price it was measured against, and an edge is published only against a price that is current.
Is the model's accuracy public?
Yes. Every NFL and NBA prediction is graded as each game finishes on the public results pages, with the season's data in CSV and JSON files, and every call is sealed with a timestamped fingerprint before the start so it cannot be edited afterward.
Example
One game in play from GET /v1/nba/edges, shortened. The numbers are illustrative.
curl https://api.lazyedge.net/v1/nba/edges -H "Authorization: Bearer YOUR_KEY"[
{
"game_id": "401910002",
"away_team": "NY",
"home_team": "TOR",
"period": 3,
"clock": "5:12",
"away_score": 74,
"home_score": 80,
"our_prob": 0.612,
"kalshi_implied": 0.57,
"edge_score": 0.042,
"pick": "TOR",
"pick_p": 0.612,
"signal_age_s": 4,
"model_ready": true
}
]