Skip to main content
The ai-coustics SDK delivers real-time speech enhancement using our Quail model family running on the AirTen inference runtime.

Models

Quail L

  • Processing latency: 30ms
  • Available sample rates:
    • 48 kHz
    • 16 kHz
    • 8 kHz

Quail S

  • Processing latency: 30ms
  • Available sample rates:
    • 48 kHz
    • 16 kHz
    • 8 kHz

Quail XS

  • Processing latency: 10ms
  • Available sample rates:
    • 48 kHz

Quail XXS

  • Processing latency: 10ms
  • Available sample rates:
    • 48 kHz
The SDK resamples the audio to the model’s sample rate as needed.

How the SDK is structured

  • Core library: C ABI delivered as static/dynamic libraries (Linux, macOS, Windows; x86_64, ARM64)
  • Language wrappers: C++, Python, and Rust packages for idiomatic usage
  • Real-time processing: 48/16/8 kHz streaming with frame-based processing and low latency

Quickstart

1

Get an SDK license

Contact us for a demo license key.
License Keys will soon be available through our developer portal.
2

Pick your wrapper

Use our language wrappers for a simpler integration.
You’re ready to process audio in real time with Quail.