Built with Laya
Laya is five days old, so nothing here is a product yet. What there is instead is the infrastructure a new model attracts when it is small enough to run on the machine already on your desk: six different runtimes, four servers that speak TypeSafe's protocol, the tooling around them — and a lot of Tetris.
Stars and last push read 2026-09-24 18 described, all with a page of their own
- GitHub
GitHub · page
NandhaKishorM/laya
Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
The upstream model, and the row the rest of this column is built around: three checkpoints, a router that picks between them per request, and a server that answers TypeSafe's own POST /v1/systemone shape. Its benchmark table puts the publisher's own measurements next to Jev numbers it has never reproduced, and says so in the line above the table.
- GitHub
GitHub · page
mizorewww/laya-mlx
Native MLX runtime for Laya typed decision models — 7–14 ms short decisions on M3 Max. No text generation, PyTorch, or cloud API.
The port that made the numbers small: 13.4 ms median for a short English decision, 7.4 ms with the multilingual checkpoint, no PyTorch in the runtime. The README is unusually careful — it says in the caption that the latency figure is a one-question benchmark and not the frame time of the Snake loop in the GIF above it.
- GitHub
GitHub · page
mizorewww/laya-coreml
Local Laya typed decisions on Apple Core ML and Neural Engine. Validated ports, ~5 ms short decisions on M3 Max, reproducible speed and energy benchmarks.
The same author's Core ML port, and the rare demo that reports energy instead of speed: 2.78× better whole-system energy per decision than compiled MLX at 4.98 ms p50. The benchmark document also says which measurements exclude terminal rendering and which do not, which is the detail that decides whether a number means anything.
- GitHub
GitHub · page
receptron/laya
Run Laya, the open-source Jev-compatible System-1 decision model, from Node.js / TypeScript via ONNX Runtime
The Node path: ONNX Runtime, no Python at runtime, and about 1.7 GB of fp32 weights cached on first call. The README claims the output matches the Python reference to four decimal places — which is exactly the kind of claim a reader should re-run on their own inputs before trusting it.
- GitHub
GitHub · page
Heman10x-NGU/openJev-verdict-2.0
Calibrated 151M Non-Autoregressive Decision Engine beating TypeSafe Jev & Laya on LocalLLaMA/typed-decisions (77.10% acc, 0.0636 Brier, 0.0144 ECE)
A 151M decision engine that claims to beat both Jev and Laya, with an in-browser WebGPU path so it can run without a server at all. It is the loudest set of numbers in this column and it comes with the least context, so the row is here as a claim to check rather than a result to quote.
- GitHub
GitHub · page
wdobry/laya-playground
A website, two games, a benchmark and an agent skill for Laya, the open-source decision model. Runs on your machine.
A website, three games, a benchmark and an agent skill in one repository. The public site has no model behind it and replays recorded runs, which is a more honest arrangement than a demo that quietly falls back to an API — and the same pages drive the real model once you clone it.
- GitHub
GitHub · page
ipenywis/laya-ultrafast
Same as jev-ultrafast but using Laya
Browser Use's ultrafast agent with the decision call swapped for local Laya, and the README is scrupulous about the split: the agent, the DOM snapshot, the executor, the safety checks and the design are theirs, only the decisions changed hands. Median about 33 ms on an M1 Max, and no per-step API cost.
- GitHub
GitHub · page
virajbhartiya/laya-vs-jev
Laya vs Jev: local MLX and hosted AI decisions playing T-Rex side by side, with live metrics and replay recording
Two models, one Chrome dinosaur, and a replay of every crash. A deterministic planner labels the moves and the models choose between them, so the README says outright that an assisted score measures the whole system — the sentence most comparison demos leave out.
- GitHub
GitHub · page
lkarlslund/laya.cpp
C++ inference for Laya typed decisions - supports CUDA, Vulkan, Core ML, CPU
ggml underneath, with CUDA, Vulkan and Core ML backends, a Jev-compatible HTTP server and request batching. It is the first port here that treats Windows as a first-class target — signed binaries rather than a build script and a paragraph of apologies, which for a C++ inference engine is the hard part.
- GitHub
GitHub · page
openlayer-ai/jevals
Agent evals and guardrails as Jev decisions: one request per trace, a fraction of a cent, fast enough for the agent loop. Runs locally with Kev or Laya.
Evals and guardrails where the judge is a decision model instead of an LLM, published by a company that already sells evals. The argument is arithmetic rather than vibes: all the evals for one trace go out as one request, for a few thousandths of a cent, which is what makes running them on every trace affordable.
- GitHub
GitHub · page
ThinkFlowLab/system1-agents
System 1 decision models (Jev, Laya, Cua-S1) as brain for agents: Browser use, computer use, games and robotics
Prebuilt agents — browser, computer, robotics, games — with a decision model as the brain and a choice of Jev, Laya or Cua-S1 underneath. There is a video at the top of the README, which is the fastest way to see what the pattern is for if the prose has not landed yet.
- GitHub
GitHub · page
angel291592/Intent-Router
Intent compiler for AI agents — converges vague requests into typed IntentSpec contracts (probe, ask, or halt before routing), the input layer for routers and typed-decision models like Jev & Laya
It starts a step earlier than the decision: turning "add caching to the user API" into a typed IntentSpec, probing for what it can look up and asking for what it cannot, then refusing to emit while the request is still underspecified. The halt is the part most routers leave out and the part that makes the output trustworthy.
- GitHub
GitHub · page
1Panel-dev/laya-server
A self-hosted API and web interface for Laya’s structured decision models, compatible with the TypeSafe Jev API format.
Docker-first, with a web interface, from a team that already ships server tooling. The part worth noticing is not the UI: it is that a project with release tags and a Chinese-language README has picked Laya up and packaged it as a service you start with one command.
- GitHub
GitHub · page
alvarobartt/sys1
System One compatible API for open decision models, e.g. Laya, written in Rust.
A Rust server for the System One protocol: candle underneath, dynamic batching by token count, CPU, CUDA and Metal backends. The headline number comes with its hardware attached — about 14 ms per query on an RTX Pro 6000 — which is more than most rows in this column manage.
- GitHub
GitHub · page
PromptEngineer48/laya-vs-jev-arena
Laya (open source, local) vs TypeSafe Jev (API): two AI models race in Snake and fight in a Mortal-Kombat-style arena. Every move is a real model decision.
The arena from the Prompt Engineer 48 video: a snake race and a fighting game, every move a real decision, and a docker one-liner that lets a person take the place of one side. A fighting game is a more searching test than it looks — it needs a decision under time pressure where the opponent has already moved.
- GitHub
GitHub · page
0xBakeer/arbiter
Serve typed-decision (System 1) models — Laya or your own — on NVIDIA GPUs or Apple Silicon, with a Jev-compatible API and coding-agent integrations
Serve Laya or a head you trained yourself, on a GPU or a Mac, behind a Jev-compatible API. It arrives with a build log rather than a pitch — a Sunday-morning post about a cat and a local Jev — and the diagram spends its space on where the model sits next to an LLM instead of on benchmark bars.
- GitHub
GitHub · page
bladedevoff/stuntd
Local proxy that learns your app's typed LLM decisions and answers them with a Laya head. Jev and OpenAI compatible.
A local proxy that records the typed decisions your application already sends to an LLM, learns them, and answers later ones with a Laya head — so most calls never leave the machine. The interesting half is not the 90% claim, it is that the training data is your own traffic rather than a benchmark.
- GitHub
GitHub · page
afshinm/laya-mps
Run Jev-style typed decisions locally on your Mac with low RAM usage and fast responses
PyTorch's Metal backend instead of MLX or Core ML, and the row that treats memory as the constraint rather than latency: about 2.1 GiB in the fast mode, and a slower 0.74 GiB mode for machines that cannot spare it. The Pong demo renders the response time on the frame, so the number in the README is visible while it runs.
What this column does not have
A verdict on which of these is good, and a star count that means what you think it means. Stars order the list and nothing else — a row with eleven stars here is the only browser build of the model in any language, and a row with six thousand is a port by somebody who got there first. Where we read the code the note says so; where we did not, it says that.
Source: GitHub search API Numbers are the authors'; the sentence is ours Built something with Laya? Send it over