Scores an explicit list of supplier ↔ reference room-name pairs. Each pair gets a calibrated match score and binary label. Use when you already know which pairs to evaluate (QA, A/B tests, custom pipelines)—unlike Room Match many-to-many, which scores the full grid and ranks per supplier.

Key Parameters

• threshold: Decision cutoff on calibrated score. Default: 0.5
• batchSize: Informational batch hint on the request. Default: 64
• lowercase: Lowercase room text before scoring. Default: true
• includeLogitRaw: Include logitRaw per result when true. Default: false
• pairs: Array of pairs to score, each with id, supplierRoomName, and referenceRoomName

Per-pair Result

scoreCalibrated, predictedLabel (1 = match at threshold), decisionThreshold, optional logitRaw.

To experiment with this API, use the sandbox key:
e2R4t6Y8u0i3O5p7A9s1D3f5G7h9J2k4