Skip to main content
Acepeak

Home / CPaaS / Media streaming

Feature · Media streaming

Your calls, live in your app.

Stream live call audio to your app, get the words as they are spoken, and act in the moment.

Live audioReal-time wordsWebhooksYour network
One call, streamed live
Live call
Caller
Agent
Stream
Live transcript
Hi, I need to move my delivery to Friday.
Done. Friday it is.
intent: reschedule
Sentiment
Webhook sent to your app
00:17

Illustrative

A recording tells you what happened. A live stream lets you change it while the caller is still on the line.

Media streaming sends call audio to your app as the call happens, so transcripts, AI agents and alerts can react in the moment.

What you build

One stream, three jobs.

The same live audio can feed a transcript, an AI agent or a monitoring rule.

Get words as they are spoken over a live stream, for real-time apps.

  • Hi, I need to move my delivery to Friday.
  • Speaker labels and timestamps
Streaming transcription

Words arrive as they’re spoken.

No waiting for the call to end. Real-time apps get each word over the live stream.

Hi,IneedtomovemydeliverytoFridayafternoon,andcouldyoutextmewhenitisontheway?

Speaker labelsTimestampsStreams live
Barge-in

Let callers interrupt.

With the audio streaming live, an AI agent hears the caller the moment they speak and adjusts its reply in real time.

  • Natural, two-way conversation
  • The agent stops talking when the caller does
  • No awkward overlap or waiting
Agent
Caller
0sCaller interruptsAgent adjusts
React live

When it happens, do something.

Set a rule on intent, sentiment or language, and the stream triggers the action while the call is still going.

When
intent == "cancel_subscription"
Then
POST webhook → retention flow
When
sentiment < -0.6 for 20s
Then
Alert lead · transfer to senior
When
caller.lang != agent.lang
Then
Translate both ways, live

Illustrative rules

Your network
Caller
In
Your switch
Streams
Your app
Receives
TLSSRTPMedia stays inside
Your network

The audio never leaves home.

Media stays on your network, so latency and margin both stay with you. Use TLS and SRTP to protect signaling and media.

  • Runs through your own switch
  • Encrypted signaling and media
  • Latency and margin stay with you
Start a stream

Four steps up.

From a key in your developer portal to a call that acts on its own.

1

Authenticate

Create an API key in your branded developer portal.

2

Stream

Send live audio to the speech endpoint.

3

Subscribe

Register webhooks for intents, sentiment shifts and finished transcripts.

4

Act

Route, alert, translate or summarise with the results.

FAQ

Media streaming, answered.

Hear the call as it happens.

Talk to partnerships about streaming live call audio from the Voice API.

Contact Us