GitHub · Runtime ports · Serving and API
receptron/laya
JavaScript services that want the decision in-process, with no Python sidecar.
What it does
Runs Laya from Node or TypeScript through ONNX Runtime, keeping the request and response shape of the Python reference and of TypeSafe's system_one call, for services whose decision-making process is already JavaScript.
The repository's own description: “Run Laya, the open-source Jev-compatible System-1 decision model, from Node.js / TypeScript via ONNX Runtime”
How it works
An ONNX session loads from a local cache that fills from Hugging Face on first use, and each call renders the state and its questions into one batch, returning a probability distribution per question.
The repository, by the numbers
Read from the GitHub API on 2026-09-24. Stars and the last push move daily — quote them with the date, the way we do.
What we checked
- the README and package page, read 2026-09-24
- the stated memory budget and the fp32 weight size
- the licence file
What we did not check. The four-decimal agreement with the Python implementation. We did not install the package or diff the two outputs.