Multi-agent framework that assigns GPT roles (PM, engineer, QA) to solve complex software tasks end-to-end.

Pure C/C++ LLM inference library — run quantized models on CPU, Metal, CUDA and more.

Official GitHub MCP server — give your AI agent full read/write access to repos, issues, PRs, and actions.

State-of-the-art sentence and text embeddings — compute semantic similarity, clustering, and dense retrieval.

Hugging Face's state-of-the-art library for diffusion-based image, video, and audio generation models.

The Natural Language Toolkit — Python's foundational NLP library for tokenization, POS tagging, parsing, and corpora.

Official TypeScript/JavaScript SDK for building MCP servers and clients — the Node.js foundation for the Model Context Protocol.

Fast and flexible image augmentation library with 70+ transforms for computer vision model training.

Fast, memory-mapped dataset library for NLP and ML — 50,000+ datasets, streaming, and Arrow-backed processing.

Conversational AI coding agent that builds entire codebases from natural-language specs with one prompt.

The largest collection of pretrained image models for PyTorch — ViT, ConvNeXt, EfficientNet, Swin, and 900+ more.

Ultra-fast tokenizer library (Rust core) — BPE, WordPiece, SentencePiece — tokenize GBs in seconds.

Run powerful open-source LLMs locally — privacy-first desktop app and Python SDK, no internet required.

Drop a screenshot or Figma design and get clean HTML/Tailwind/React/Vue code generated by GPT-4o or Claude.

Fine-tune LLMs with RLHF, PPO, DPO, SFT, and GRPO — the standard library for aligning language models.

Meta's Segment Anything Model 2 — real-time promptable segmentation for images and videos with streaming memory.

The world's most popular computer vision library with Python bindings — image processing, video, and ML pipelines.

Run PyTorch training scripts on any hardware — single GPU, multi-GPU, TPU — with minimal code changes.

Lightweight neural voice activity detector that distinguishes speech from silence/noise in real time.

CTranslate2-powered reimplementation of Whisper with up to 4× faster inference and lower memory usage.

Instant voice cloning with precise style control — tone, emotion, accent, and rhythm from a single short reference clip.

OpenAI's lightweight Python framework for building multi-agent systems with handoffs, tracing, and guardrails.

8-bit and 4-bit quantization for PyTorch — run and fine-tune LLMs on consumer GPUs with minimal quality loss.

Blazing-fast DataFrame library written in Rust with a Python API — handles datasets that don't fit in RAM.