npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@uncharted-ai/donut-runtime

v1.0.2

Published

🍩 Donut Corp Runtime — native PC inference engine, no C++ compilation needed

Downloads

489

Readme

🍩 Donut Runtime

Runtime próprio do Donut Corp — inferência local em pure JavaScript, sem compilar C++, usando todo o poder do seu PC.

Como funciona

Ao contrário do Ollama (que usa llama.cpp compilado), o Donut Runtime:

  • Lê GGUF direto — parser binário próprio em Node.js
  • Roda inferência em JS puro — V8 auto-vectoriza com AVX2/NEON automaticamente
  • Zero compilaçãonpm install e pronto, funciona em qualquer PC
  • Detecta hardware — usa todos os cores da CPU, detecta GPU automaticamente

Arquitetura

donut-runtime/
├── src/
│   ├── gguf/reader.js        ← Parser GGUF binário (sem dependências)
│   ├── engine/
│   │   ├── ops.js            ← matmul, attention, rmsNorm, RoPE (SIMD via V8)
│   │   ├── transformer.js    ← Forward pass LLaMA/Mistral/Phi/Gemma
│   │   ├── hardware.js       ← Auto-detect CPU/GPU/RAM
│   │   └── index.js          ← DonutEngine (interface principal)
│   ├── tokenizer/index.js    ← BPE tokenizer (lê vocab do GGUF)
│   ├── models/               ← Registry + download manager
│   ├── server/               ← API OpenAI-compatible
│   └── cli/                  ← CLI `donut`

Install

npm install -g @uncharted-ai/donut-runtime

Sem compilação. Funciona direto.

Uso

# Ver hardware detectado
donut hw

# Baixar modelo
donut pull tinyllama

# Chat
donut run tinyllama

# Ou apontar direto para um .gguf
donut run ~/models/meu-modelo.gguf

# API server (OpenAI-compatible)
donut serve

Como usar como biblioteca

import { DonutEngine } from '@uncharted-ai/donut-runtime';

const engine = new DonutEngine();
await engine.load('./modelo.gguf', { ctx: 2048 });

// Streaming
for await (const token of engine.chat([
  { role: 'user', content: 'Olá!' }
])) {
  process.stdout.write(token);
}

Performance

O V8 (engine do Node.js) auto-vectoriza operações em Float32Array:

  • x86_64: usa SSE4.1 / AVX2 automaticamente
  • ARM64 (Apple Silicon, etc.): usa NEON automaticamente
  • Threads: usa todos os cores menos um (configurável)

Para máxima performance em modelos grandes, use GPU via donut serve com --gpu-layers.

Modelos suportados

Qualquer modelo em formato GGUF: LLaMA 2/3, Mistral, Phi-3, Gemma 2, Qwen 2.5, TinyLlama e outros.

ATENÇÃO!

Donut runtime NÃO tem um repositorio no GitHub

License

MIT © Donut Corp / Uncharted AI