Analytics API

The same engine that powers our replays — now as a service for your platforms and products.

What the API does

  • Performance Rating (PR) and position equity
  • Compare a played move against the best alternative (equity loss)
  • Move grades: best / ok / doubtful / error / blunder
  • Backgammon and Long Narde through a single interface
  • Multi-phase neural nets — tens of milliseconds per request

Who it's for

  • Coaches running analysis on student matches
  • Game platforms shipping built-in analytics
  • Tournaments producing PR reports and ratings
  • Researchers working on board-game AI

Documentation

POST /api/v1/analyze endpoint

JSON in: position + played move + match context. JSON out: equity-loss, grade, ranked list of alternatives with win-probability breakdowns. API-key authenticated.

Request

POST https://nardex.ai/api/v1/analyze
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{
  "game_type": "narde",
  "position": {
    "pips": [
      -1,  0,  0,  0, -1,  0,   // points 1–6   (Black's starting quarter)
      -2,  0, -1,  0,  0, -8,   // points 7–12  (Black's outer field)
       1,  0,  0,  0,  1,  0,   // points 13–18 (White's outer field)
       2,  0,  1,  0,  0,  8    // points 19–24 (White's starting quarter)
    ],
    "off": [2, 2],
    "bar": [0, 0]
  },
  "decision": {
    "type": "checker_play",
    "dice": [3, 5],
    "played": [{"from": 24, "to": 21}, {"from": 21, "to": 16}]
  },
  "context": { "player": "x", "match_state": null }
}

Response

{
  "alternatives": [
    {
      "play": [{"from": 24, "to": 21}, {"from": 21, "to": 16}],
      "equity": 0.142,
      "wins": { "win": 0.71, "gammon": 0.18, "backgammon": 0.04 }
    },
    ...
  ],
  "equity_loss": 0.000,
  "grade": "best",
  "analysis_depth": "ply_zero"
}

Request access

Tell us briefly about your project, expected request volume and integration — we'll send a key and pricing tiers.