Voice and WebRTC calling

ChannelsUpdated 2026-07-28

Voice on Arino One is WebRTC-based: calls are placed and answered in the browser console, can originate from your website or from a paired Ping Me device, and land on the same contact timeline as every other channel. It is provisioned per instance rather than switched on by default, on the client's own Infobip account. Voice/WebRTC calls are billed at $0.009 per minute while the instance stays on an active $50/month Arino Core connection. Calls that reach the public telephone network also incur number rental and termination costs, paid directly to Infobip by the client, which vary by country and number type.

The $0.009/minute voice/WebRTC rate is available only while the instance stays on an active $50/month Arino Core connection. If you stop paying the monthly connection fee or disconnect from Arino Core, this rate no longer applies and you are responsible for sourcing your own voice/TURN provider.

What voice is for

Voice is the escalation channel. Most conversations should stay asynchronous — WhatsApp, SMS or web chat — because they are cheaper, they queue cleanly, and they leave a written record. Voice exists for the small share of interactions where typing is the wrong medium: a caller standing at a physical entry point, a complex complaint that has bounced twice in chat, or a high-value contact who simply prefers to talk.

Because Arino One's voice is WebRTC rather than a classic PBX, there is no desk phone, no SIP client to install and no separate telephony console. An agent who is signed into the inbox is reachable; an agent who is not, is not. That keeps presence honest and means call routing follows the same team structure as message routing.

The three ways a call starts

From the website. A call button in the chat widget opens an audio channel straight from the visitor's browser to your routing group. No number is dialled, so there is no international call cost to the visitor and no number to publish.

From a Ping Me device. A physical Ping Me button raises a call to the assigned routing group and identifies itself by device and location, so the agent sees "Loading bay — rear gate" before they answer rather than an unknown caller. Device pairing and location assignment are covered in Ping Me device setup.

From the console outbound. An agent can call a contact from the timeline. If the contact only has a phone number, this needs a PSTN leg and therefore an Infobip voice number; if the contact is on the website, it rings their open browser session.

Routing, hours and fallbacks

A routing group is the unit of configuration: a named set of users or teams, a ring strategy, a schedule and a fallback. Ring strategies are either simultaneous — everyone in the group rings and the first to answer wins — or sequential, which is useful when you want a specialist to get first refusal before the call spills to a general queue.

Fallbacks matter more than the ring strategy. An unanswered call should never simply drop. Configure voicemail so the recording is transcribed into an inbox conversation, or overflow to a second group, or send an automatic WhatsApp message to the caller acknowledging the missed call and offering a callback slot. That last pattern converts a failed call into a live async conversation and is the single highest-value voice configuration most teams miss.

Recording, disclosure and retention

Recording is disabled until you deliberately turn it on for a routing group. When enabled, attach the disclosure announcement — in most European jurisdictions callers must be told before recording begins, and a pre-connect announcement is the cleanest way to evidence that. Recordings inherit your instance's retention policy, so they expire on the same schedule as messages and are removed by an erasure request just like any other personal data. See retention policies and erasure requests for how that is enforced.

Quality and network requirements

WebRTC is sensitive to network conditions rather than bandwidth volume. A call needs roughly 100 kbps in each direction, which is trivial, but it needs low jitter and open UDP. Corporate networks that force all traffic through a proxy, or firewalls that block UDP outright, will degrade calls to the TURN relay path and add latency. If agents report choppy audio, check for UDP blocking before assuming a platform fault — it is the cause in the large majority of cases.

How voice differs by connection status

Connection statusVoice provisioningWho bills the minutes
Connected ($50/month Arino Core connection)Client's own Infobip account, with Arino support and management$0.009/min via Arino; PSTN number rental and termination paid directly to Infobip by the client
UnconnectedClient's own Infobip account, configured by the client, or the client's own voice/TURN providerNo Arino billing or the $0.009/min rate; Infobip (if used) bills the client directly

Every client owns their Infobip voice application regardless of connection status. The difference is support and rate access: on a connected instance, Arino assists with configuration and the $0.009/minute rate applies. On an unconnected instance, you configure and troubleshoot voice yourself, you handle numbering and any regulatory registration, you must source your own voice/TURN provider since the $0.009/minute rate is unavailable, and there is no Arino invoice or SLA attached to voice on that instance.

Costs to plan for

Voice and WebRTC calls cost $0.009 per minute while the instance stays on an active $50/month Arino Core connection. Calls that terminate on the public telephone network also incur number rental and termination charges, which the client pays directly to Infobip, and which vary by country and number type. If the instance is unconnected, the $0.009/minute rate does not apply and you must source your own voice/TURN provider. Ping Me adds its own charges — $0.10 per button press and $0.009 per talk minute — on top of the $200 one-time device cost. Full figures are in billing and plans and voice and WebRTC billing.

Steps at a glance

  1. Enable the voice channelAsk Arino to activate voice on your instance. Voice is not provisioned by default — it is enabled per instance because it needs a WebRTC application and, if you want PSTN legs, numbers on your Infobip account.
  2. Assign numbers or stay browser-onlyDecide whether calls are browser-to-browser (web calling from your site or a Ping Me device) or need real phone numbers. Browser-only needs no numbering; PSTN calls need Infobip voice numbers in each market you call from or to.
  3. Create a routing groupIn Settings → Voice, create a routing group and add the users or teams who should ring. Set the ring strategy (all at once or in order) and a fallback for unanswered calls — voicemail, a message to the inbox, or an overflow group.
  4. Set opening hours and out-of-hours behaviourAttach a schedule to the routing group. Outside hours, send callers to voicemail so the recording lands in the inbox as a conversation rather than ringing an empty desk.
  5. Configure recording and retentionTurn recording on per routing group, add the disclosure announcement your jurisdiction requires, and set the retention window so recordings are deleted on schedule alongside your other data.
  6. Test a call end to endPlace a call from the web caller and from a paired Ping Me device, confirm it rings the right group, answer it in the console, then check that the call appears on the contact's timeline with its duration and recording.

Frequently asked

What does voice cost?

Voice and WebRTC calls are billed at $0.009 per minute while the instance stays on an active $50/month Arino Core connection. Calls that reach the public telephone network also incur number rental and termination costs paid directly to Infobip by the client, which vary by country and number type. See billing and plans for the full rate card.

Do I need phone numbers to use voice?

No. Browser-to-browser calls — from your website's call button or a Ping Me device to an agent in the console — work without any numbering. You only need Infobip voice numbers if calls must terminate on the public telephone network.

Where do calls appear afterwards?

Every call, answered or missed, is written to the contact's timeline alongside their WhatsApp, SMS and email history, with duration, the agent who handled it, and the recording if recording was enabled.

Does voice work the same whether or not I'm connected to Arino Core?

The functionality is identical either way, since every client provisions voice on their own Infobip account. On a connected instance, the $0.009/minute WebRTC/voice rate applies and Arino provides support. On an unconnected instance, that rate does not apply — you must source your own voice/TURN provider — and Infobip bills you directly for any PSTN usage; Arino neither provisions nor invoices it.

Is call recording GDPR-safe by default?

Recording is off until you enable it. When you do, enable the disclosure announcement and set a retention window — recordings are personal data and are covered by the same erasure and retention tooling as messages.

Next step

Apply this to your own deployment

This guide describes decisions we make on live instances. Tell us your channels, systems and region and we will map it to an architecture outline, a provisioning plan and an indicative commercial model — usually within one business day.