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

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

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

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

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

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

Production-grade Stable Diffusion studio with a node-based workflow editor, ControlNet, IP-Adapter, and full REST API.

Meta AI's promptable image segmentation model that can segment any object from a single click or bounding box.

Apache-licensed 7B language model from Mistral AI — beats Llama 2 13B on most benchmarks at half the size.

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

Hugging Face Transformers: state-of-the-art pre-trained models for NLP, vision, audio, and multimodal tasks.

Facebook AI's library for efficient text classification and word representation learning — trains in seconds on millions of documents.

Resemble AI's state-of-the-art open-source TTS model with voice cloning, emotion exaggeration, and zero-shot speaker adaptation.

Battle-tested deep learning TTS toolkit with XTTS, VITS, and 1000+ pretrained voices across 20+ languages.

Ship an entire LLM as a single self-contained executable — one file, every OS, no install needed.

Meta AI's modular object detection platform supporting Mask R-CNN, Faster R-CNN, DETR, and panoptic segmentation.

Microsoft's MIT-licensed 3.8B SLM — instruction-tuned, runs on CPU/mobile, punches far above its weight class.

82M-parameter Apache-2.0 text-to-speech model with high naturalness and multiple voice styles.

OpenAI's open-source speech recognition model — accurate multilingual transcription and translation in one model.

Alibaba's Apache-2.0 7B instruction model — top multilingual performance, 128k context, strong at coding and math.

Self-hosted, open-source AI coding assistant — code completion, chat, and codebase indexing on your own infra.

Suno's transformer-based text-to-audio model that generates realistic speech, music, and sound effects from text.

BERT pre-trained on 12GB of diverse English legal text for superior legal NLP: case law, contracts, legislation, and EU/US court documents.

Meta's self-supervised ViT model producing universal visual features for classification, segmentation, depth estimation, and retrieval.