Overview

Lead scoring turns behaviour into a number. You define rules — each a signal, an operator, a value and a point award — and the API sums the ones that match a record, producing a score and an A–D grade with a per-rule breakdown so you can see exactly why.

The timeline is the other half of the same question: a single merged feed of everything that has happened to a contact or company.

Authentication

Authorization: Bearer cm_your_api_key

Lead scoring

Signals

Signals are a closed allowlist per entity type — an unknown signal is 422, never a silently-false rule.

Contact

Company

Operators

eq, neq, gt, gte, lt, lte, contains, is_set, is_not_set — but which ones are legal depends on the signal's kind:

is_set and is_not_set take no value — one sent is dropped.

Grades

The rule object

{
  "id": "77aa…",
  "tenant_id": "a0b1…",
  "entity_type": "contact",
  "name": "Opted in on WhatsApp",
  "signal": "whatsapp_opt_in",
  "operator": "eq",
  "value": true,
  "points": 20,
  "is_active": true,
  "created_at": "2026-08-06T09:00:00Z",
  "updated_at": "2026-08-06T09:00:00Z"
}

GET /api/v1/crm/lead-scores/rules

Oldest first — the order you built them in.

POST /api/v1/crm/lead-scores/rules

curl -X POST https://api.callmissed.com/api/v1/crm/lead-scores/rules \
  -H "Authorization: Bearer cm_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "entity_type": "contact",
    "name": "Stale — no contact in 30 days",
    "signal": "days_since_last_conversation",
    "operator": "gt",
    "value": 30,
    "points": -15
  }'

Errors name the exact problem, for example operator 'contains' is not valid for the boolean signal 'has_email'. Allowed: eq, neq, is_set, is_not_set.

PATCH / DELETE /api/v1/crm/lead-scores/rules/{rule_id}

PATCH takes every field except entity_type, all optional. The signal / operator / value triple is re-validated against the merged result, so you cannot change the signal in one call and leave an illegal operator behind.

Reading scores

GET /api/v1/crm/lead-scores

Highest score first — the ranked list.

GET /api/v1/crm/lead-scores/{entity_type}/{entity_id}

{
  "id": "88bb…",
  "tenant_id": "a0b1…",
  "entity_type": "contact",
  "entity_id": "4411…",
  "score": 62,
  "grade": "B",
  "breakdown": [
    { "rule_id": "77aa…", "name": "Opted in on WhatsApp", "signal": "whatsapp_opt_in", "operator": "eq", "value": true, "points": 20 },
    { "rule_id": "99cc…", "name": "Has an open deal", "signal": "deal_count", "operator": "gte", "value": 1, "points": 42 }
  ],
  "computed_at": "2026-08-17T05:00:00Z"
}

breakdown lists only the rules that matched, which is what makes a score explainable to a salesperson.

404 No score has been computed for this record when the record has never been scored — call recompute first.

POST /api/v1/crm/lead-scores/recompute

{ "entity_type": "contact", "requested": 3, "computed": 3 }

Idempotent, and ids are de-duplicated before the 200 cap is checked. If any id is not yours you get 404 2 of 50 contact ids were not found in this tenant; nothing was scored — nothing is partially computed.

Scores are not recomputed automatically after you change a rule. Re-run the affected records yourself.

Timeline

GET /api/v1/crm/timeline

A merged, newest-first feed of everything attached to one record.

curl "https://api.callmissed.com/api/v1/crm/timeline?entity_type=contact&entity_id=4411…&types=note,task" \
  -H "Authorization: Bearer cm_your_api_key"

[
  {
    "id": "note:aa22…",
    "type": "note",
    "occurred_at": "2026-08-16T12:00:00Z",
    "title": "Note added",
    "summary": "Renewal call went well — wants a Hindi voice agent.",
    "actor": "Ravi K",
    "ref_id": "aa22…",
    "meta": {}
  }
]

An unknown or another tenant's entity_id returns an empty array, not a 404. The timeline never confirms whether a record exists — do not use it as an existence check.

Read-only: there is no write scope and no way to post to a timeline. It is assembled from the underlying records.

Errors

Nothing on this page consumes credits.