Decibri
Monthly visits
2,183
2026-10
Monthly growth +50%

Decibri - Real-time audio I/O for Python, Node.js, and Rust

Decibri is an audio library for real-time voice applications. It captures microphone audio, plays sound through system speakers, and provides optional local voice activity detection. Python and Node.js packages include prebuilt binaries; Rust is compiled from source. Audio can be streamed to files, WebSockets, speech-to-text services, or voice agents.

Visit WebsiteUpdate this product

0Views

Capabilities

Real-time microphone capture

Captures microphone audio and outputs streamable audio chunks. Configure sample rate, channels, buffer frames, device selection, and int16 or float32 output; the default chunk is 100 ms.

System speaker playback

Plays audio through the operating system's speaker and accepts streamed audio as input. The Speaker API can be used as a write target alongside capture for full-duplex audio.

Local voice activity detection

Detects speech and silence and outputs voice activity events. The optional bundled Silero VAD v6.2 model runs locally in Rust through ONNX Runtime without requiring a cloud API.

Audio streaming and recording

Streams audio through Python iterators, Node.js Readable streams, or Rust channels, and can write recordings to files. Pipe the output to files, WebSockets, speech services, or voice agents.

Browser audio support

Provides browser audio capture through the Node.js package's AudioWorklet implementation. Conditional exports select the browser implementation when the package is bundled for a browser.

Pricing

Decibri is presented as an installable software library, and the supplied pages provide no paid plan or price to compare. Install it through the Python, Node.js, or Rust package ecosystem.

Use cases

Stream microphone audio to speech recognition

When an application needs live speech transcription, capture microphone audio and stream its PCM output to a speech-to-text service. Decibri provides language-native streams and integration guides for listed services.

Record microphone audio to a file

When an application needs a saved recording or audio for batch processing, capture from a microphone and write the stream to a file. The record_to_file() helper supports setting a sample rate for the target format.

Provide audio I/O for a voice agent

When building a real-time voice agent or assistant, capture microphone input, stream it to the agent, and play audio through system speakers. The APIs support streaming and full-duplex audio workflows.

Detect speech locally in an audio stream

When an application needs speech and silence events without a cloud detection API, enable the bundled Silero VAD. The model runs locally through ONNX Runtime.

Limitations

  • The Rust crate is compiled from source with cargo; the prebuilt-binary installation described for Python and Node.js does not apply to Rust.

  • The Windows ARM64 Python package requires Python 3.11 or later; other listed Python platforms support Python 3.10 and later.

  • The Node.js package requires Node.js 18 or later.

  • The listed macOS prebuilt binaries are for Apple Silicon; the page does not list an Intel macOS binary.

Who it is for

  • Developers building real-time voice applications who need microphone and speaker audio I/O across programming languages.

  • Developers connecting live audio to speech recognition, wake-word detection, WebSocket pipelines, or voice agents.

  • Python, Node.js, and Rust developers who need a consistent audio interface across Windows, macOS, and Linux.

FAQ

What audio formats can Decibri output?

It outputs signed 16-bit little-endian PCM by default, and 32-bit float output is also available. The format can be configured for audio-processing requirements.

Does Decibri work in a browser?

Yes. The Node.js package includes browser support, and conditional exports select an AudioWorklet implementation when bundled for a browser.

Do Python and Node.js installations need build tools?

No. Their packages install prebuilt binaries for supported platforms. The Rust crate is compiled from source by cargo.

Which Node.js versions does Decibri support?

Node.js 18 and later are supported.

Which Python versions does Decibri support?

Python 3.10 and later are supported on Windows, macOS, and Linux, except Windows ARM64, which requires Python 3.11 or later.

What license is Decibri released under?

Decibri is released under the Apache-2.0 license.

Comments

Please sign in to leave a comment.
No comments yet. Be the first to share your thoughts!