Skip to main content

​
Start here

Use POST /v1/score to score a company or deal and get a deterministic response with structured reasons. Scoring is idempotent for the same normalized input and algorithm version.

​
What scoring returns

Each scoring response includes:
  • score: integer from 0 to 100
  • band: low, medium, good, or very_good
  • dimensions: weighted scoring dimensions
  • reasons: structured supporting reasons
  • explanation: compact natural-language summary
  • algorithm_version, input_hash, and output_hash for deterministic snapshots

​
Score a company

curl -X POST "https://api.sanka.com/v1/score" \
  -H "Authorization: Bearer <access_token>" \
  -H "Content-Type: application/json" \
  -d '{
    "object_type": "company",
    "record_id": "11111111-2222-3333-4444-555555555555"
  }'
Company scoring combines firmographic fit and record quality signals.

​
Score a deal

curl -X POST "https://api.sanka.com/v1/score" \
  -H "Authorization: Bearer <access_token>" \
  -H "Content-Type: application/json" \
  -d '{
    "object_type": "deal",
    "record_id": "66666666-7777-8888-9999-000000000000"
  }'
Deal scoring combines qualification hygiene, engagement recency, pipeline progression, and associated company quality.

​
Interpret the result

Typical response shape:
{
  "data": {
    "object_type": "deal",
    "record_id": "<record_uuid>",
    "snapshot_id": "<snapshot_uuid>",
    "algorithm_key": "deal_score_v1",
    "algorithm_version": "v1",
    "input_hash": "<sha256>",
    "output_hash": "<sha256>",
    "score": 78,
    "band": "good",
    "dimensions": [],
    "reasons": [],
    "explanation": "Strong qualification hygiene with room to improve activity freshness."
  },
  "message": "ok",
  "ctx_id": "<ctx_id>"
}
Use the API playground in API Reference → Data Scoring to test both company and deal requests.