Laya projects and examples: 106 builds on GitHub, X and YouTube

Showing 1 to 48 of 106

Introducing Clef: our open-source decision models, and new RL fine-tuning platform
Cloudflare's launch post for Clef and Clef-flash, two decision models on Workers AI that accept the Jev request format and open weights under Apache 2.0. It covers the architecture, Cloudflare's own benchmark tables against Jev, Kev and Laya, and a new reinforcement learning service.
Michelle Chen (@michellechen)140k viewsOpenIntroducing Clef: our open-source decision models, and new RL fine-tuning platform on blog.cloudflare.com
ArticleSupport inbox triage
Decision Model Leaderboard
Cloudflare's live leaderboard for decision models on the Jev Decision Index suite, with Clef, Clef-flash, Jev, Kev-9B and Laya plotted by score and latency. Cloudflare built and ran it, so treat the ranking as vendor-run.
Cloudflare (@ritakozlov)41k viewsOpenDecision Model Leaderboard on clef-evals.workers-ai-mle.workers.dev
Article
NandhaKishorM/laya
Decision model from Convai Innovations that fully fine-tunes an encoder and adds a typed-decision head: a 421M English checkpoint on ModernBERT-large and a 322M multilingual one on mmBERT-base, with a router picking between them by detected language. Returns choice, score and noul answers, and reports expected calibration error of 0.081 after temperature fitting, down from 0.466 as shipped.
NandhaKishorM21k starsOpenNandhaKishorM/laya on github.com
Alternative
Laya-MLX
An independent MLX port of Laya for Apple Silicon, 13.4ms median for a short English decision and 7.4ms on the multilingual checkpoint, with a Snake demo that plays at up to 75 moves per second under 1GB of memory.
mizorewwwOpenLaya-MLX on github.com
6.1k starsShort decision 13.4 ms P50Snake demo 75.4 moves/s
ToolReal-time control
Laya-CoreML
Core ML ports of all three Laya checkpoints for Apple Silicon, reaching 4.98ms P50 on the Neural Engine for a short multilingual decision, with a Snake demo sustaining about 50 decisions per second.
mizorewwwOpenLaya-CoreML on github.com
1.4k starsShort decision 4.98 ms P50Snake demo 49-50 dec/s
ToolReal-time control
@receptron/laya
Node.js and TypeScript package that runs the Laya decision model through ONNX Runtime, so JavaScript projects can call it without Python.
receptronOpen@receptron/laya on github.com
374 stars3-question call ~140 msModel size ~1.7 GB
Tool
FluidUse
A Swift package for local computer use on Apple Silicon that reads a form through the Accessibility API and asks a small on-device model what belongs in each field. Its Laya integration runs FluidInference's Core ML port of Laya on the Neural Engine at 3.7 ms per decision on an M5 Pro, about seven times faster than the upstream GPU figure.
FluidInferenceOpenFluidUse on github.com
152 starsPer decision 3.7msSpeedup ~7x faster
ProjectTyped tool dispatch
Laya Ultrafast
A local, open-weight port of browser-use's jev-ultrafast agent that swaps Jev's hosted decisions for Laya running on Apple Silicon through MLX, while keeping the original hosted Jev policy as a fallback mode.
ipenywisOpenLaya Ultrafast on github.com
145 starsDecision latency ~33ms
ProjectTyped tool dispatch
Laya vs Jev: T-Rex Race
Laya, running locally on Apple Silicon via MLX, and Jev, over TypeSafe's hosted API, play Chrome's T-Rex game side by side on a shared obstacle course, with response times, survival streaks and crash replays recorded.
virajbhartiya93 starsOpenLaya vs Jev: T-Rex Race on github.com
ProjectReal-time control
laya.cpp
Standalone C++ inference for Laya's typed decisions using ggml, with CUDA and Vulkan backends, a Jev-compatible /v1/systemone HTTP server, and automatic request batching.
lkarlslundOpenlaya.cpp on github.com
85 starsCUDA BF16 speedup 1.11-2.71x
Tool
jevals
A Python evals and guardrails library for agent traces that uses Jev-style decision models instead of an LLM judge, running Jev through TypeSafe or Vercel, or Kev or Laya locally on a Mac. All checks for a trace go out as one request in a few hundred milliseconds.
openlayer-aiOpenjevals on github.com
81 starsVariance vs LLM 92-913x
ToolLLM guardrails
Chat Seek
A VS Code extension that finds past Claude Code, Codex and OpenCode conversations from a plain-language description, using a local Laya model to rerank the keyword search results.
fstandhartinger67 starsOpenChat Seek on github.com
ProjectSemantic reranking
codejunkie99/keel
Local-first macOS coding workspace built in Rust with a local Laya selector and optional hosted Jev, routing fresh coding tasks and validating each chosen route on the host.
Avid (@Av1dlive)Opencodejunkie99/keel on github.com
41 starsCost/time cut ~80%
ProjectAgent routing and skill selection+1
Laya vs Jev: AI vs AI Arena
Laya, running locally, and Jev, over TypeSafe's hosted API, face off in a snake race and a Mortal-Kombat-style fight, both answering the same typed questions on the same input; from a Prompt Engineer 48 video.
PromptEngineer4829 starsOpenLaya vs Jev: AI vs AI Arena on github.com
ProjectReal-time control
0xBakeer/arbiter
Serves typed-decision models, including Laya or your own, behind a Jev-compatible API on NVIDIA GPUs or Apple Silicon, with coding-agent integrations.
0xBakeer23 starsOpen0xBakeer/arbiter on github.com
Tool
Laya MPS
A local server running Laya's typed-decisions checkpoint on a Mac's GPU through PyTorch MPS, with a Pong demo and a memory-versus-latency tradeoff reaching about 32ms median decisions at 2.1GiB RAM on an M5 Pro.
afshinm18 starsOpenLaya MPS on github.com
ToolReal-time control
laya_ex
A native Elixir runtime for Laya built on Nx and Bumblebee, downloading the official checkpoint on first load and answering typed choice and noul questions in-process.
ChristianAlexander15 starsOpenlaya_ex on github.com
Tool
Laya Drift
An OpenCode plugin that embeds the running session state with Laya and scores how far it has drifted from the calibrated plan using weighted Jensen-Shannon divergence, shown live as a sidebar score.
pythongiant14 starsOpenLaya Drift on github.com
ToolSelf-consistency checks
Laya Jev Lab
An independent benchmark of Jev against Laya on 40 Chinese support-ticket classifications, plus a confidence-gated cascade that calls Laya first and escalates to Jev, matching Jev's 78% accuracy at 1.8x the speed.
yibie9 starsOpenLaya Jev Lab on github.com
ProjectConfidence-gated actions+1
Lev
Laya inference ported to Jolt, running Clojure on Chez Scheme, with a confidence gate that escalates uncertain answers from the Laya encoder to a GGUF chat model scored by token log probabilities.
jlt-commons8 starsOpenLev on github.com
ToolConfidence-gated actions
localdecide
A local browser-agent decider running Laya's open weights, speaking the same /v1/systemone wire format as Jev so existing jev-ultrafast tooling works by changing one base URL. Ships a head-to-head benchmark against hosted Jev.
ChenneyZhuang7 starsOpenlocaldecide on github.com
ProjectReal-time control
Laya Rust Inference (candle)
Pure-Rust inference for Laya built on candle, with CPU, Metal and CUDA backends, no Python, torch or ONNX export step, and weights pulled straight from the Apache-2.0 checkpoint.
aovestdipaperino7 starsOpenLaya Rust Inference (candle) on github.com
Tool
kevala
Runs Laya and Kev decision models inside the browser: a zero-dependency Rust engine compiled to WebAssembly with WebGPU kernels, downloading a pinned int8 model pack and answering choice, score and noul questions with no server or API key.
bvolpato6 starsOpenkevala on github.com
Tool
Laya4j
A pure-Java 17+ SDK that ports the Python Laya decision framework into a deployable inference library, running on Apache 2.0 multilingual checkpoints via ONNX Runtime and the Hugging Face Hub.
githubMJ6 starsOpenLaya4j on github.com
Tool
jevbench
A reproducible benchmark running Jev, Laya and four other classifiers over three public datasets, reporting accuracy, calibration, latency, throughput and cost side by side. Jev answers one choice question per example through the Decisions API.
dhruvmehra5 starsOpenjevbench on github.com
Project
Laya Plays Super Mario Bros. 3
Unmodified Laya clears Super Mario Bros. 3 World 1-1 on an NVIDIA DGX Spark through Stable Retro NES emulation, with 16.79ms median inference and every frame's decision recorded for replay.
cv5 starsOpenLaya Plays Super Mario Bros. 3 on github.com
ProjectReal-time control
sysone-bench
Runs Laya and Jev on 751 byte-identical questions across nine suites; Jev sweeps triage, guardrails and moderation, Laya wins agnews and mnli, and both miscalibrate on emotion.
instax-dutta4 starsOpensysone-bench on github.com
Project
laya-zig
A dependency-free Zig CPU runtime that reads Laya's safetensors weights directly and runs the full forward pass, tokenizer included, in one binary; marked work in progress, with a CLI, browser UI and a Snake probe.
li-ming14 starsOpenlaya-zig on github.com
Tool
Laya Model Router (glukicov)
A FastAPI request router that uses Laya as a System 1 decision layer to pick between small, medium and powerful downstream models, matching GPT-5 nano's routing accuracy at far lower latency in the author's evaluation.
glukicov4 starsOpenLaya Model Router (glukicov) on github.com
ProjectIntent and model routing
laya-candle
Native Rust inference for Laya using the Candle framework, with fused Metal kernels and tiled attention tuned for M1 Pro, loading Laya's original safetensors with no Python runtime.
b0xtch4 starsOpenlaya-candle on github.com
Tool
baronunread/leanest
A local-first test selector that asks one semantic relevance judgment per test to decide which e2e tests a code change can safely skip, failing open to the full suite on low confidence. Judge is pluggable between classifier.dev, Jev over HTTP, or self-hosted Laya via ONNX Runtime. Ships a bundled GitHub Action.
baronunread4 starsOpenbaronunread/leanest on github.com
ToolConfidence-gated actions
@r4ai/laya-web
A client-side inference runtime for Laya-MLX decision models via ONNX Runtime Web, with WebGPU acceleration and automatic Wasm SIMD fallback, runnable in a Web Worker to keep the UI thread responsive.
r4ai3 starsOpen@r4ai/laya-web on github.com
Tool
chunklaya
Measures how Laya's decision head stops attending to input past roughly 200 tokens, then fixes it by chunking documents into passages and screening each one. Tested on 400 documents, and against Jev on a public needle benchmark.
myxamediyar3 starsOpenchunklaya on github.com
ProjectRAG passage filtering
laya-pong
A browser pong game whose left paddle is decided by one Laya choice question per frame, 18.7ms p50 on Metal, while the right paddle plays fixed arithmetic for comparison.
aovestdipaperino3 starsOpenlaya-pong on github.com
ProjectReal-time control
Laya Snake: Local Decision Experiment
A desktop Pygame Snake game with HUMAN, Laya AI and greedy-baseline modes, showing a live decision dashboard with real SDK probability bars, entropy confidence and per-move inference time.
AmirTahaMim2 starsOpenLaya Snake: Local Decision Experiment on github.com
ProjectReal-time control
ComfyUI Sequential GPU Router
Two ComfyUI nodes that free VRAM between models on a single GPU: one unloads cached image models, and a Laya-based router classifies the enhanced prompt's style before an image model loads. Measured on a 12 GB RTX 3080 Ti.
highdeftant2 starsOpenComfyUI Sequential GPU Router on github.com
Project
Laya Router Skill
A cross-platform Agent Skill and MCP tool that gives Codex, Claude Code, OpenCode and Pi local Laya decision routing, with explicit invocation verified end to end on Apple Silicon and cross-platform CI.
wangmiaozero2 starsOpenLaya Router Skill on github.com
ToolAgent routing and skill selection
Laya for Home Assistant
A fully local Home Assistant Assist conversation agent that scores registered intents and exposed entities with Laya in one forward pass, dispatching simple commands directly and escalating compound or low-confidence requests to a fallback LLM.
allenporter2 starsOpenLaya for Home Assistant on github.com
ProjectIntent and model routing
laya-mcp (wsargent)
A stdio MCP server built on the laya-mlx runtime, exposing preset tools for support triage, prompt-injection guarding, moderation and email classification, plus a generic typed choice, score and noul tool.
wsargent2 starsOpenlaya-mcp (wsargent) on github.com
ToolTyped tool dispatch+2
laya-needle
A fork of the Needle Chrome extension that finds webpage passages by meaning, swapping its hosted Jev calls for Laya running locally so no API key is needed.
50bbx1 starsOpenlaya-needle on github.com
ProjectRAG passage filtering
hermes-skill-router
A Hermes plugin that pre-routes each task to a skill before the main model sees it, classifying against a roughly 300-skill index with SemIf by default or a legacy local Laya model. Measured 67 percent top-1 for SemIf against 56 percent for Laya on the same pools.
cdepuy1 starsOpenhermes-skill-router on github.com
ToolAgent routing and skill selection
pydecide
Unified Python client for typed decisions with a fallback chain across Jev, OpenRouter, Laya and cross-encoder models.
gopaljigaur1 starsOpenpydecide on github.com
ToolIntent and model routing
Feishu Message Classification Benchmark
A frozen 64-case benchmark of Feishu-style workplace messages, scoring Jev and Laya on a single choice question and on a four-question workflow. Jev scored 100 and 98.44 percent correct, Laya scored 31.25 and 28.12 percent, on the same requests.
Adkid-Zephyr1 starsOpenFeishu Message Classification Benchmark on github.com
ProjectSupport inbox triage
zero-shot-ie-bench
Compares GLiNER 2.5, GLiFormer, GLiClass, Laya, Von and Jev on zero-shot information extraction and classification in one web UI.
umstek1 starsOpenzero-shot-ie-bench on github.com
ProjectStructured extraction
docker-laya
A dockerized FastAPI service that loads Laya's English, multilingual and typed-decisions checkpoints behind a router, and serves choice, score and noul questions over HTTP with API-key or Basic auth.
chneau1 starsOpendocker-laya on github.com
Tool
jeffrey
A coding-agent CLI that splits work between a decision model, Jev by default or self-hosted Laya, which only picks the next tool and scores progress and risk, and an OpenAI-compatible LLM that fills in the tool arguments.
thomasbrueggemann1 starsOpenjeffrey on github.com
ProjectTyped tool dispatch
Jev vs Open Decision Models
Zero-shot classification benchmark across four datasets: Jev leads decisively on topic tasks, and Laya's advertised edge over Jev does not reproduce, tying on emotion and trailing elsewhere.
elcronos1 starsOpenJev vs Open Decision Models on github.com
Project
LayaKit
A Swift package that runs one Laya typed decision per call on-device via Core ML, ported from laya-coreml, with no Python needed at runtime.
tahby1 starsOpenLayaKit on github.com
Tool

All entries