Score a record with one POST.

1. Mint a key. In the app, open Settings, API keys, then Create key. The secret (ax1m_sk_...) is shown once. Save it. API scoring is available on Pro and above.
2. Grab a model_id. Any trained model in your workspace. It looks like mdl_abc123.
3. POST a record. Send the fields you have. You do not need every field the model trained on.

curl -X POST https://api.ax1om.ai/v1/score \
  -H "Authorization: Bearer ax1m_sk_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model_id": "mdl_abc123",
    "records": [
      { "title": "VP Sales", "industry": "SaaS", "employees": 250 }
    ]
  }'

You get back a fit score, a conversion likelihood, and your month-to-date usage:

{
  "model_id": "mdl_abc123",
  "scores": [
    { "score": 82, "conversion_likelihood": 0.823, "enriched": false }
  ],
  "records_scored": 1,
  "cache_hits": 0,
  "usage": { "current": 151, "limit": 50000, "remaining": 49849, "in_grace": false, "backstop": 100000 }
}

Every call sends your key as a Bearer token. The webhook endpoint also accepts an X-API-Key header for CRM tools that map headers more easily. Keys start with ax1m_sk_. Treat the secret like a password: it is shown once at creation and stored only as a hash.

Authorization: Bearer ax1m_sk_your_key

curl -X POST https://api.ax1om.ai/v1/score \
  -H "Authorization: Bearer ax1m_sk_test_your_test_key" \
  -H "Content-Type: application/json" \
  -d '{ "model_id": "mdl_abc123", "records": [ { "title": "VP Sales" } ] }'

Request body

Response

Score row (each item in scores)

Parameters

Send the field values as a flat JSON body:

curl -X POST "https://api.ax1om.ai/v1/webhooks/score?model_id=mdl_abc123" \
  -H "X-API-Key: ax1m_sk_your_key" \
  -H "Content-Type: application/json" \
  -d '{ "Title": "VP Sales", "Industry": "SaaS", "LeadSource": "Webinar" }'

Response

{
  "score": 87,
  "tier": "hot",
  "percentile": 87,
  "conversion_likelihood": 0.8732,
  "top_reason_1": "VP title (2.3x lift)",
  "top_reason_2": "SaaS industry",
  "top_reason_3": "Webinar source"
}

When you score a record that is not yet in your existing scored data, ax1om scores it from exactly the fields you send. Nothing is invented and no account matching happens on the request path. The response tells you plainly with the enriched flag:

The score is always real, never a placeholder. enriched: false is a signal about input completeness, not score quality.

Every /v1 error returns the same shape. Branch on error.code, never on the HTTP status or the message text. The codes are frozen: new ones may be added, existing ones never change.

{
  "error": {
    "code": "invalid_api_key",
    "message": "API key required. Send Authorization: Bearer ax1m_sk_...",
    "docs_url": "https://ax1om.ai/docs/api"
  }
}

A per-key rate limit (default 10 requests/second) guards against bursts and returns rate_limited when tripped. A separate monthly metered quota counts the records you score. The monthly limit is a soft cap: live keys on a paid plan keep scoring past it in a grace band (in_grace: true, remaining goes negative) and only hard-stop with quota_exceeded at the backstop. Test keys meter into their own free bucket with a hard cap.

Every batch response includes a usage object so you always know where you stand this month:

Batch up to 1000 records per call to stay well under the rate limit.

The API is versioned by URL prefix (/v1). Additive changes, new response fields, new optional request fields, and new endpoints ship within /v1 without a version bump. Breaking changes, renaming or removing a field, changing a type, or changing an error code, get a new prefix, never a silent change to /v1.