Audio API
Unbody leverages Mux (opens in a new tab) to provide audio hosting and streaming services. With Mux you can expect a reliable audio streaming experience with many more features to enhance your audio content.
Core Capabilities
- Streaming: Deliver your audios to any device. Use Mux's multi-platform player or your preferred HLS player.
- Automatic Transcription: Generate captions for your audio files (English only).
- Static Renditions: Access static renditions of your audio files for offline viewing, direct downloads and other use-cases.
Getting Started
Prerequisites
Before you begin, make sure you have the following:
- An active Unbody account.
- Access to the Unbody GraphQL endpoint.
Fetch Your Audio
Every audio in Unbody is associated with a Mux asset. To stream an audio with Mux player, you need the playbackId
associated with the asset, or you can use the hlsUrl
directly with your preferred HLS player.
query {
Get {
AudioFile {
playbackId
hlsUrl
}
}
}
Setup Your Player
Here's an example of how you can use Mux React Player to stream your audio:
import MuxAudio from "@mux/mux-audio-react";
const AudioPlayer = ({ title, assetId, playbackId }) => {
return (
<div>
<MuxAudio
controls
playbackId={playbackId}
streamType="on-demand"
style={{ height: "100%", maxWidth: "100%" }}
/>
</div>
);
};
Please refer to Mux Documentation (opens in a new tab) for more guides and examples.
Auto Generated Captions (English Only)
Mux uses OpenAI's Whisper model to automatically generate captions for your audios - read more (opens in a new tab).
Unbody stores these captions as SubtitleFile
and SubtitleEntry
objects, here's how you can retrieve them:
query {
Get {
SubtitleFile(
where: {
operator: Equal
path: ["media", "AudioFile", "originalName"]
valueText: "your-audio-file.mp3"
}
) {
entries {
... on SubtitleEntry {
start
end
text
}
}
}
}
}
Here are some examples of how you can leverage the auto-generated captions:
- Perform semantic search to find audio files or specific parts of audios based on your query.
- Use the
ask
operator to get answers based on the audio transcription. - Use the
generate
operator to perform RAG (Retrieval-Augmented Generation) on audio transcription.
Furthermore, if you have the auto-generated fields feature enabled in your project, Unbody utilizes the transcriptions to generate summaries, extract keywords, named entities, and more. These additional fields are added to both the AudioFile
and SubtitleFile
objects.
Static Renditions
Mux generates static renditions of assets for offline viewing, direct downloads, scenarios where streaming HLS streaming is not possible, and other use cases. For audio files, Mux produces a single .m4a
file. Here's how you can access it:
query {
Get {
AudioFile {
files {
url
name
ext
size
bitrate
}
}
}
}
User Limits
Free Plan:
- Audio files longer than 10 minutes are not supported.
- Audio files larger than 60MB are not supported.
- A maximum of 15 minutes of audio in total can be stored at once.
For an extended limit, consider upgrading your plan or contact support.
Paid Plan:
- Audio files larger than 2GB are not currently supported.
Processing audio files doesn't impact build time; billing is based on duration and streaming time. See pricing details here.
Known Issues and Limitations
- Auto-generated captions are currently available exclusively for English audio content.
- Accessing playback metrics is not currently supported.
- Direct access to the Mux API is not currently supported.