Infinity is a high-throughput, low-latency serving engine for text-embeddings, reranking models, clip, clap and colpali

🤖🤮 Hate the yuck codes that agents generated? Try `alint`, a ESLint like toolchain for intent driven code check, freeze your skills, AGENTS.md to lint rules

Language model tokenization at GB/s

Bitwarden-compatible server running on Cloudflare Workers

将被语言阻隔的作品，带到读者的语言中。Bringing literature into your language.

Policy-driven, layered isolation and containment

A tutorial on modern GPU programming for machine learning systems

🎡 A configurable git server written in Node.js

Sudoless Asitop

Lightweight Tracy profiler integration for Python

Pure-Rust LLM runtime: one binary serves (OpenAI-compatible), runs local agents, and distills models on their own rollouts — on Apple Silicon and NVIDIA. No Python on the hot path.

A tmux sidebar that monitors AI coding agents across all windows and sessions in real-time

mock operating system, a la macintosh 1984, built in react

Reimplementing Git in Rust. Like a boss.

C++ ggml runtime hub for multilingual ASR and TTS models: Cohere Transcribe, Parakeet TDT, Voxtral, Canary 1B v2, etc, plus universal forced alignment, and more

⬆️ Media Capture in Swift

cuJSON: A Highly Parallel JSON Parser for GPUs

迄今最完美的 Apple Music 歌词体验还原 · macOS & Windows · 逐词同步 · Skia GPU 加速 Pixel-perfect Apple Music visuals -- A universal Now Playing lyrics overlay for macOS & Windows — works with Spotify, Apple Music, a…

Magenta RealTime 2: An Open-Weights Live Music Model

Optocam Zero is a Raspberry Pi Zero based compact digital camera made using off the shelf components.

Unicode spinner animations as raw frame data

agent multiplexer that lives in your terminal.

Tile-Based Runtime for Ultra-Low-Latency LLM Inference

AI-native SwiftUI component library with full-stack recipes — connect via MCP for instant access.

Write dynamic binary analysis tools in Python

Pure-Python RISC-V instruction generator with built-in functional coverage, auto-regression, and CI-ready dashboards.

Quantization, kernels, runtime and inference engine for mobiles, wearables, smart home and robots.

Foundation model for tiny devices; 14mb, 26m params, 1-6k toks/sec on mobiles, wearables smart home and robots.