Skip to content
acoustics-agent

acoustics-agent

Legacy Physics meets Modern AI — Blazingly fast underwater acoustic engine powered by Numba.
Get Started View on GitHub

Key Features

Section titled “Key Features”

High Performance

Powered by the pyacoustics engine with Numba JIT acceleration, reaching near-Fortran efficiency with 1000+ rays/sec.

AI-Native Design

Native support for LLM tool use via standardized Skills, enabling AI agents to drive simulations through natural language.

Proven Accuracy

Validated against 27+ classic acoustic benchmarks (Munk, Pekeris, RAP), ensuring parity with legacy toolbox results.

Modern Workflow

Declarative YAML configuration for human-readable scenario management, perfectly suited for modern dev pipelines.

Install from the repository README

Section titled “Install from the repository README”

Clone the GitHub repository and enter the source directory. Then follow the README by choosing the standard Anaconda environment or a local editable pip installation:

Terminal window
git clone https://github.com/weifengsd/acoustics-agent.git
cd acoustics-agent


# Recommended: standard Anaconda environment
conda install numpy scipy numba pyyaml matplotlib


# Or install the current source checkout with pip
pip install -e .

For full parity with legacy Bellhop/Kraken, install Acoustics Toolbox separately and point AT_BIN_PATH to its at/bin directory:

Terminal window
export AT_BIN_PATH=/path/to/at/bin

Perspective

Section titled “Perspective”

“Legacy Physics meets Modern AI.”

acoustics-agent is more than just a solver; it’s acoustic infrastructure designed for the AI era. Whether building autonomous underwater systems or conducting deep-sea research, it provides the agile computation you need.