Core conceptsCalls

Make a call

Trigger a single AI call from your application
View as Markdown

The Calls API is the main reason to integrate with HollrAI programmatically.

Build and deploy your agent in app.hollr.ai, then use your backend to trigger a call whenever something happens in your product: a new lead arrives, an order needs confirmation, a candidate needs screening, a customer needs a follow-up, or any other workflow that benefits from a single phone conversation.

Use /v1/call/make-call to trigger one outbound call.

$curl -X POST https://api.hollr.ai/v1/call/make-call \
>  -H "hollr-api-key: $HOLLR_API_KEY" \
>  -H "Content-Type: application/json" \
>  -d '{
>    "campaign_id": "CAMPAIGN_ID",
>    "to": "+15551234567",
>    "metadata": {
>      "orderId": "ORD-4821",
>      "customerId": "CUST-1042"
>    }
>  }'

campaign_id and to are required. metadata is optional and accepts arbitrary JSON values you want associated with the call request.

When to use this vs. a batch

  • Single event-driven call: use make-call.
  • Multiple contacts, one request, optionally scheduled for later: use batches.
  • Repeated or recurring workflows: keep the business logic in your application and trigger the appropriate endpoint when the call should happen.

HollrAI handles the voice conversation itself. Your application decides when to call and who to call.

Response

make-call currently passes through the response returned by HollrAI’s underlying call server, so the API reference intentionally exposes the response as an open object rather than pretending it has a fixed schema.

Limits

make-call is subject to the standard 60 requests/minute limit per API key, plus a 5 concurrent-call limit per API key.

See errors and rate limits for the full details.

After triggering a call

A completed call can be read through the Conversations API. Use the conversation record when your application needs the transcript, summary, sentiment, recording information, call identifiers, or other call data.