CricketBrain
Open Source AGPL-3.0Neuromorphic pattern recognition inspired by cricket hearing
Description
CricketBrain is a biomorphic inference engine in Rust that processes temporal patterns via delay-line coincidence detection — with no matrix multiplication and no GPU. Inspired by the auditory system of the field cricket, it reaches TPR=1.0 and FPR=0.0 across all tested SNR levels. The core is no_std, making it suitable for embedded devices. Language bindings exist for Rust, C/C++, Python, and WebAssembly.
Preview
Platforms
Linux Windows macOS
Replaces the following proprietary tools
TensorFlow Lite → CricketBrain
PyTorch Mobile → CricketBrain
Self-Hosting
CricketBrain can be self-hosted on your own infrastructure. Visit the official documentation for installation instructions.
Documentation →Categories
Tags
#neuromorphic
#rust
#embedded
#no-std
#signal-processing
#stdp
#spiking-neural-network
Similar Tools
Ollama
Run large language models locally with ease
MIT ● Easy Self-hostable
ChatGPT Midjourney DALL-E
Last: Feb 2026 105.0k
View details →
LocalAI
Local OpenAI-compatible API for various AI models
MIT ●● Medium Self-hostable
ChatGPT Midjourney DALL-E
Last: Feb 2026 26.0k
View details →
Stable Diffusion WebUI
Powerful user interface for AI image generation
AGPL-3.0 ●● Medium Self-hostable
ChatGPT Midjourney DALL-E
Last: Feb 2026 145.0k
View details →