Outputs

Profiles, claims, intent scores, and sample responses partners receive from the SDK and API.

After events are accepted and archived, partners read intelligence back through the profile APIs. V1 returns resolved profiles, claims, and intent scores.

What a partner can read today

Output	Endpoint	Scope	Timing
Resolved profile	GET /v1/profiles/:customer_id	profile:read	Synchronous
Underlying claims	GET /v1/profiles/:customer_id/claims	claims:read	Synchronous
Intent scores	GET /v1/profiles/:customer_id/intents	intents:read	Synchronous
Event status	GET /v1/events/:event_id	events:read	Synchronous
Effective consent	GET /v1/customers/:id/effective-consent	consent:read	Synchronous
Workflow status	GET /v1/workflows/:workflow_id	workflows:read	Synchronous

Pass consent_purpose on profile reads when you need the consent gate enforced for that call. If the purpose is not granted, the API returns 403 consent_required.

Resolved profile

Swapnice does not overwrite contradictory facts. Each attribute is resolved from claims using highest_confidence_then_most_recent, with alternatives retained.

curl "https://api.sandbox.swapnice.com/v1/profiles/cust_8f2a1c?consent_purpose=personalization" \
  -H "Authorization: Bearer $ACCESS_TOKEN"

{
  "customer_id": "cust_8f2a1c",
  "resolved_at": "2026-08-26T16:10:02.000Z",
  "facts": [
    {
      "attribute": "favorite_series",
      "resolved_value": "yugioh",
      "confidence": 0.91,
      "resolution_method": "highest_confidence_then_most_recent",
      "consent_purpose": "personalization",
      "evidence": [
        {
          "id": "clm_101",
          "customer_id": "cust_8f2a1c",
          "attribute": "favorite_series",
          "value": "yugioh",
          "source": "partner_app",
          "observed_at": "2026-08-20T18:00:00.000Z",
          "context": "profile_selection",
          "consent_purpose": "personalization",
          "confidence": 0.91,
          "provenance": {
            "source_partner": "partner_app",
            "source_system": "ios",
            "collected_at": "2026-08-20T18:00:00.000Z"
          }
        }
      ],
      "alternatives": [
        {
          "value": "pokemon",
          "confidence": 0.64,
          "reason": "lower_confidence_or_older_claim"
        }
      ]
    },
    {
      "attribute": "recent_purchase_object",
      "resolved_value": "sku_blue_eyes_tin",
      "confidence": 1,
      "resolution_method": "highest_confidence_then_most_recent",
      "consent_purpose": "personalization",
      "evidence": [],
      "alternatives": []
    }
  ]
}

What you can do with this:

• Personalize the partner app from facts the user actually granted
• Show provenance when a value is inferred versus explicitly selected
• Keep a defensible audit trail instead of a single overwritten profile field

Claims

Claims are the unresolved evidence behind a profile. Use them when you need the full history, not just the winner.

{
  "data": [
    {
      "id": "clm_101",
      "customer_id": "cust_8f2a1c",
      "attribute": "favorite_series",
      "value": "yugioh",
      "source": "partner_app",
      "observed_at": "2026-08-20T18:00:00.000Z",
      "context": "profile_selection",
      "consent_purpose": "personalization",
      "confidence": 0.91,
      "provenance": {
        "source_partner": "partner_app",
        "source_system": "ios",
        "collected_at": "2026-08-20T18:00:00.000Z"
      }
    }
  ],
  "pagination": { "has_more": false, "next_cursor": null }
}

Intent scores

Intents are grouped into five categories.

Category	Meaning	Example objects
outcome	Intent to purchase, participate, or complete a collection	SKU, event, set
sentiment	Positive or negative preference	Product, brand, experience
affiliate	Self-label or group association	Team, fandom, community
habit	Repeated interest or activity	Series, category, venue
activation	Readiness for the next step in the partner product	Onboarding step, offer

{
  "data": [
    {
      "id": "evt_9c21",
      "category": "outcome",
      "object": "sku_blue_eyes_tin",
      "confidence": 1,
      "consent_purpose": "personalization",
      "evidence": [
        {
          "id": "clm_220",
          "customer_id": "cust_8f2a1c",
          "attribute": "purchase.completed",
          "value": "sku_blue_eyes_tin",
          "source": "partner_app",
          "observed_at": "2026-08-26T16:04:00.000Z",
          "consent_purpose": "personalization",
          "confidence": 1,
          "provenance": {
            "source_partner": "partner_app",
            "source_system": "ios",
            "collected_at": "2026-08-26T16:04:00.000Z"
          }
        }
      ]
    },
    {
      "id": "evt_9c18",
      "category": "habit",
      "object": "yugioh",
      "confidence": 0.72,
      "consent_purpose": "personalization",
      "evidence": []
    }
  ],
  "pagination": { "has_more": false, "next_cursor": null }
}

In V1, an intent row is derived from archived events for that customer. Use high-confidence habit and outcome rows to decide what to show next in the partner product.

Event acceptance and status

Immediately after ingest:

{
  "event_id": "evt_9c21",
  "status": "accepted",
  "sequence": 7,
  "accepted_at": "2026-08-26T16:04:00.120Z",
  "processing_status": "queued"
}

Later, GET /v1/events/evt_9c21:

{
  "event_id": "evt_9c21",
  "customer_id": "cust_8f2a1c",
  "status": "processed",
  "accepted_at": "2026-08-26T16:04:00.120Z",
  "processed_at": "2026-08-26T16:04:01.004Z"
}

Statuses: accepted, processing, processed, failed, rejected.

Related

• Events and intent ontology
• What a pilot enables
• Example use cases
• API overview