ggml-org/llama.cppThe C/C++ inference engine for LLMs built on the GGML tensor library, letting quantized models run on CPUs, Apple Silicon and consumer GPUs. Much of the local-LLM ecosystem, Ollama included, sits on top of it.
zed-industries/zedA high-performance, multiplayer code editor written in Rust from the creators of Atom, using its own GPU-accelerated UI framework (GPUI) and increasingly positioned as an AI-native editor.
astral-sh/uvAn extremely fast Python package and project manager written in Rust, from Astral (the Ruff team). It sets out to replace pip, pip-tools, pipx, poetry, pyenv and virtualenv behind a single tool.
ollama/ollamaA tool to run open large language models (Llama, Mistral, Gemma, Qwen, DeepSeek, gpt-oss and more) locally with a single command, bundling model download, quantization and a REST plus CLI interface over a llama.cpp-style engine.
duckdb/duckdbAn in-process analytical (OLAP) SQL database — often called 'SQLite for analytics' — that runs embedded inside your application or notebook and queries Parquet, CSV and Arrow directly with no server to manage.
browser-use/browser-useA Python library that lets AI agents control a real web browser — reading the page and taking actions through Playwright — so an LLM can complete tasks on sites that have no API.
projectdiscovery/nucleiA fast, template-driven vulnerability scanner: security teams write YAML templates describing how to detect a CVE or misconfiguration, and Nuclei runs them at scale across hosts and web apps.
oven-sh/bunAn all-in-one JavaScript and TypeScript toolkit — runtime, bundler, transpiler, test runner and npm-compatible package manager — built on JavaScriptCore and pitched as a faster drop-in for Node.js.
honojs/honoA small, ultrafast web framework written in TypeScript that runs on any JavaScript runtime — Cloudflare Workers, Deno, Bun, Node, AWS Lambda — with a familiar Express-like API.