Repos

Convert → quantize → deploy path that turns Laya (and Kev, NanoJev, PlayJev) into an INT8 ONNX service on the official wire protocol. Runtime is onnxruntime, tokenizers, numpy.

Open-source project LaoA-GrokBot: 一个可定制、可随机组合并生成分享卡的 GrokBot 表情动作实验室。. Extends or complements Grok Bot for builders.

SGLang patches and launchers delivering very high single-card throughput on a workstation GPU.

A native macOS pet that shows Muse's live status, accepts typed or spoken requests and receives replies in several formats, so Muse feels present on the desktop.

Elixir library that downloads the official checkpoint on first load and answers typed choice and noul questions; pick EMLX on Apple Silicon or EXLA for CPU/NVIDIA.

An LLM engine written from scratch in HIP for AMD Strix Halo and Qwen3.8-Flash-Next, reaching about 50 tokens per second out. Built to see how far a single-model engine can go.

No Python, no torch, no ONNX export. Loads the ungated safetensors directly and runs any of the three checkpoints on CPU, Metal or CUDA.

Plugin giving you personal pixel-art bots with their own instructions, in the spirit of Grok Bot and Hermes Bots.

Tests whether the open Laya model on Apple's MLX can stand in for a paid judge, scoring model outputs locally with typed answers.

Source code behind a music video made with Opus 5.5 inside the Cursor editor.

A llama.cpp fork that speeds ROCm prefill and fits multi-token prediction for GSQ quantizations on AMD gfx1151 hardware.

Setup and prompt for screenshot/keyboard control of GTA Vice City.

One hundred self-contained HTML studies, from generative art to physics toys and interfaces, with a gallery index, screenshots and the prompts.

Open-source project growth-grok-bots: Growth-team bots for Grok Bot. Extends or complements Grok Bot for builders.

Two LLM routers behind one endpoint compared on 180 labelled requests: Laya picks small, medium or powerful in one forward pass versus a GPT-5 nano reasoning router. Local and free for the Laya side.

IQ3_XXS 177B build served from a 24GB laptop GPU with 64GB RAM.

Open-source project grokbot-linux-port: . Extends or complements Grok Bot for builders.

Same-brief agentic build in which Astra High and Gemini 3.8 Flash High each produce a playable 3D AI-civilization simulation, with outputs to compare.

Decompile of the Russian space strategy game reproducing original binaries by Astra.

Open-source project GrokBotDev: An open, agent-run directory of Grok Bot prompts, plugins, and use cases. PRs are the write API.. Extends or complements Grok Bot for builders.

Measured run using 6.6GiB of VRAM and about 48GiB of RAM on a laptop-class card.

Open-source project guaca: A desktop workspace (inspired by Grok Bot & Muse) for AI agents that collaborate, write code, and use tools. Run locally. Extends or complements Grok Bot for builders.