Stream Translation with Palabra Broadcaster Web Service

This guide walks you through setting up a broadcaster using the Palabra web interface. The setup lets you stream live audio or video to Palabra for real-time translation and multilingual streaming for your events.

General Flow

• Create a Palabra account to access Broadcaster.
• Create a Stream and configure:
  • input protocol (how Palabra receives your original stream)
  • output protocol (how Palabra restreams the translated version)
  • source language(s) for your original stream
  • target languages to translate into (multiple languages in parallel are supported)
  • optional settings to adjust your setup
• Send your stream to the Palabra server (or Palabra pulls it from your server, depending on the protocol).
• Palabra translates your stream in real time.
• Palabra restreams the translated version to your server or to a web player with low latency.

Accessing Broadcaster

1. Open the Palabra web portal
2. Log in to your Palabra account
3. Navigate to the Broadcaster section

---

Broadcaster setup process

1. Click Create Stream.
2. Enter a name for your stream.
3. Select the Input type (protocol).
   • Choose RTMP or SRT if you're streaming from OBS Studio or similar software.
   • Choose Microphone or Microphone + Camera if you want to stream directly from your device (great for offline events or quick live setups).
4. Choose the languages for the broadcast:
   • Speaking: the language the speakers will use.
     • Choose one language, or enable "Multiple source languages" to specify several source languages if your original stream will include multiple languages.
   • Translate to: the languages your stream will be translated into. The number of target languages is limited by your tariff.
5. Click Create Stream.

After clicking Create Stream, you will see the Stream settings page.

---

Stream settings

The stream settings are divided into three sections (tabs): Stream routing, Translation, and Audio. Each section is described below.

Stream routing

The Stream routing section consists of one Input and one or more Outputs.

• Input:
  • You the input type you have chosen on previous step.
  • You cannot delete the input, but you can change it (for example, switch to another source).
• Outputs:
  • You have one "Captions" output by default.
  • You can add new outputs, edit existing ones, or remove them.

INPUT CONFIGURATION

Input configuration depends on the input type you chose when creating the stream.

Input type	Description	Configuration
RTMP	Displays the RTMP URL to which you should send your original stream.	No configuration in Palabra. Point your encoder to this URL.
SRT	Displays the SRT URL to which you should send your original stream.	No configuration in Palabra. Point your encoder to this URL.
Microphone	Lets you select a microphone and start streaming directly from the Palabra web interface. Audio only.	Select a microphone and click "Start streaming", when you are ready.
Microphone + Camera	Lets you select a microphone and camera and start streaming from the Palabra web interface.	Select devices and click "Start streaming", when you are ready.

Notes for RTMP and SRT inputs:

• You can switch the URL layout by clicking the gear icon in the top-right corner. Disabling the Single URL field setting splits the URL into separate fields (useful for some software).
• As soon as you start streaming to Palabra, the stream status changes to Active and translation begins automatically.

How to change the input type:

• Click the gear icon in the top-right corner of the "Input" block on the "Stream routing" tab.
• Choose a different input type from the dropdown.
• Click Apply changes in the bottom bar.

OUTPUTS CONFIGURATION

The Outputs section lets you configure the destination and format of the translated stream.

You start with a default "Captions" output, and you can add more outputs to get the translated stream in different formats in parallel.

Available outputs:

Output type	Description	Configuration
Captions	Provides a link to a page that shows original and translated captions in real time, and can optionally play audio for a chosen language.	No additional configuration. Use the provided link. You can share this link with anyone, it is available without authorization.
RTMP	Provides fields where Palabra restreams the translated RTMP to your server. One input per target language to stream each translation as a separate RTMP stream.	Fill in your server URL (with Stream Key included). Or disable "Single URL field" in the block's menu to enter URL and Stream Key separately.
SRT	Provides a single field where Palabra restreams the translated SRT to your server. All target-language tracks are multiplexed into the same stream, so there is only one input (unlike RTMP).	Fill in your server URL (with Port and Stream Key included). Or disable "Single URL field" in the block's menu to enter URL, Port, and Stream Key separately.

Captions with Audio

If you want to hear the audio translation on the Captions page:

• enable the "Captions with audio" toggle inside the "Captions" output block
• click Apply changes
• use the additional link that contains Captions with Audio.

You can share this link with anyone - it’s accessible without authorization.

How To add a new output

1. Click the Add output button in the top-right corner of the Outputs section.
2. Select the desired output type.
3. Fill in the required fields.
4. Click Apply Changes to save it.

---

Translation

In this section, you can add, edit, or remove stream languages.

After making any changes, click Apply changes to save.

Note: The number of target languages is limited by your tariff.

---

Audio

In this section, you can configure your stream's audio settings:

• Sync translated audio: Keep translated audio synchronized with the original by slightly shifting the original stream to reduce latency between original and translation.
• Translation delay: Adjust the timing offset (in seconds) used for "Sync translated audio".
• Original base volume during translation: Background volume of the original track when the translation track is playing but no translated speech is active.
• Original ducked volume during translation: Background volume of the original track when the translation track is playing and translated speech is active.
• Voice: Select the voice used to translate your speech. You can optionally use your own cloned voice for a more personalized result.
• Multiple speaker voices: Use different voices for speakers with high and low voice timbre.

For more details, see the Voices section.

---

Starting a Stream

To start streaming:

1. Make sure you have at least one output.
2. Fill in the required fields.
3. Click Apply changes (if shown) to save.
4. Then do one of the following:

• Click Start streaming in the Input section if your input is "Microphone" or "Microphone + Camera".
• Or start sending your external stream to Palabra if your input is "RTMP" or "SRT".

Translation and restreaming start automatically within 5–30 seconds.