API reference

Automations

Start an automation from your own system — an order, a signup, a form — and track what happened next. Build the flow in the MYWA app; the API decides who goes through it and when.

List automations

GET/automations

Requires automations:read. Returns every automation you can start, with how many people it has enrolled and when it last ran. Archived automations aren’t listed.

bash

curl "$MYWA_API/automations" -H "Authorization: Bearer $MYWA_KEY"

json

{
  "automations": [
    {
      "id": "wf_9f2",
      "name": "Order follow-up",
      "description": "Thank the customer, then ask for a review after delivery",
      "status": "ACTIVE",
      "trigger": "From another app",
      "enrolled": 412,
      "lastRunAt": "2026-08-27T09:14:02.000Z"
    }
  ]
}

Start an automation for a contact

POST/automations/:id/enroll

Requires automations:write. Identify the person by phone or contactId — one of them is required.

Body parameters

Field	Type	Description
`phone`	`string`	E.164 phone number. A number that is new to your workspace creates the contact. Required unless you send contactId.
`contactId`	`string`	An existing contact id. Required unless you send phone.
`name`	`string`	Contact display name.
`attributes`	`object`	Key→value pairs (string, number, or boolean). Merged into the contact, not replaced.
`vars`	`object`	Data for this run only — not saved on the contact. Steps read it as {{vars.hook.<key>}}.

Unknown fields are rejected with 400 validation_error, so a typo never silently drops data.

curl -X POST "$MYWA_API/automations/wf_9f2/enroll" \
  -H "Authorization: Bearer $MYWA_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "phone": "+15551234567",
    "name": "Jane Doe",
    "attributes": { "city": "Austin" },
    "vars": { "orderId": "A-1049", "amount": 49.99 }
  }'

Returns 202 Accepted — the run is queued, not finished. contactCreated tells you whether the phone number was new.

json

{ "runId": "run_7c1", "contactId": "cts_31a…", "contactCreated": true }

Using vars in your steps

Anything under vars is available to the automation’s steps. With the request above, a message step can say Thanks for order {{vars.hook.orderId}}!.

The automation must be live

It has to be published and have an active trigger. Otherwise the request fails with 409 not_published or 409 no_trigger. Your workspace’s guardrails (re-entry, opt-outs, plan limits) still apply — see automation errors for which ones are worth retrying.

List runs

GET/automations/:id/runs

Requires automations:read. Test runs from the builder aren’t included.

Query parameters

Field	Type	Description
`status`	`string`	One of PENDING, RUNNING, WAITING_REPLY, SLEEPING, WAITING_EVENT, COMPLETED, FAILED, CANCELLED.
`contactId`	`string`	Only runs for this contact.
`limit`	`number`	1–100. Defaults to 25.
`cursor`	`string`	The nextCursor from the previous page.

bash

curl "$MYWA_API/automations/wf_9f2/runs?status=COMPLETED&limit=25" \
  -H "Authorization: Bearer $MYWA_KEY"

json

{
  "runs": [ { "id": "run_7c1", "status": "COMPLETED", "…": "…" } ],
  "nextCursor": "run_6b0…"
}

Pass cursor=<nextCursor> to fetch the next page. nextCursor is null on the last page.

Get one run

GET/automation-runs/:runId

Requires automations:read. Fetch a run by the runId you got when you started it — no need to know which automation it belongs to.

bash

curl "$MYWA_API/automation-runs/run_7c1" -H "Authorization: Bearer $MYWA_KEY"

json

{
  "id": "run_7c1",
  "workflowId": "wf_9f2",
  "contactId": "cts_31a…",
  "conversationId": "cnv_88b…",
  "status": "COMPLETED",
  "currentNodeId": null,
  "error": null,
  "exitReason": "end_node",
  "goal": {
    "name": "Left a review",
    "reachedAt": "2026-08-27T09:31:44.000Z",
    "value": null,
    "currency": null
  },
  "startedAt": "2026-08-27T09:14:02.000Z",
  "finishedAt": "2026-08-27T09:31:45.000Z"
}

Field	Type	Description
`status`	`string`	Where the run is now: still going (PENDING, RUNNING), paused (WAITING_REPLY, SLEEPING, WAITING_EVENT), or done (COMPLETED, FAILED, CANCELLED).
`currentNodeId`	`string \| null`	The step the run is on. null once it has finished.
`error`	`string \| null`	Why the run failed. null unless status is FAILED.
`exitReason`	`string \| null`	How a finished run ended, e.g. end_node when it reached the end of the flow.
`goal`	`object \| null`	null until the run passes a "Goal reached" step; then the goal name, when it was reached, and an optional value and currency.
`startedAt`	`string`	ISO-8601 time the run started.
`finishedAt`	`string \| null`	ISO-8601 time the run ended. null while it is still going.

const run = await mywa.get(`/automation-runs/${runId}`);

if (run.goal) console.log('goal reached:', run.goal.name);
else if (run.status === 'FAILED') console.error('run failed:', run.error);
else console.log('status:', run.status);

See it in a real flow

The Use cases guide starts an order follow-up from a checkout and connects a website form through Zapier.