Developer seeks testers for browser-based decentralized AI network

A developer has published a browser-based distributed AI inference project and is actively recruiting testers. The system uses WebAssembly (WASM) or pure JavaScript depending on device capabilities to allow web browser users to contribute computing power to AI inference tasks. The architecture distributes matrix multiplication — a fundamental operation in neural networks — across multiple participating devices; as more users join, the mathematical workload is shared more widely, enabling faster computation of neural network layers.

The primary goals are to validate whether this decentralized approach can function at scale, to measure its efficiency, and to quantify bandwidth requirements. Users can visit https://ecthqmainserver.orfe-climb.ts.net/ to begin testing. The interface includes a button to disable contribution to allow viewing progress without actively participating.

The developer has been transparent about limitations and current state. The AI model powering the proof-of-concept is intentionally basic and not meant to deliver high-quality results; the focus is on validating the infrastructure rather than the model's performance. The developer expects bugs during this testing phase and has committed to fixing them quickly. Known issues include connections dropping after a delay, which may be resolved by switching networks. A wiki documenting the project is planned for the near future. This remains an early-stage experiment in distributed, browser-based AI inference.