A PyTorch framework for generative recommendation that uses an RQ-VAE to map items to semantic IDs and a Transformer to predict the next item.

An AI-powered penetration testing automation framework that orchestrates security tools and multi-agent AI to perform adaptive security assessments and evidence-backed reporting.

An open-source LLM framework for extracting structured data and insights from documents with precise source references and automatic justifications.

A training-time alignment framework for reference-to-video generation that improves identity consistency and reference fidelity without adding inference-time overhead.

A minimalist framework for training video world models based on diffusion-forcing, used for future video prediction, 3D reconstruction, and robotic planning.

Matcha-TTS is a fast, non-autoregressive text-to-speech architecture that uses conditional flow matching to enable high-quality, efficient speech synthesis.

A high-performance, scalable LLM library written in JAX and Python for pre-training and post-training large models on Google Cloud TPUs and GPUs.

An interface between Hugging Face libraries and OpenVINO that accelerates AI model inference and optimization on Intel architectures.

A tool that converts trained scikit-learn and Keras models into portable C99 code for efficient inference on microcontrollers and embedded systems.

A modern C++ library for high-performance GPU inference of CNN models using NVIDIA TensorRT, featuring a safe engine cache and zero-copy Python bindings.

A collection of modern, TypeScript‑typed JavaScript packages (`@huggingface/hub`, `@huggingface/inference`, `@huggingface/mcp-client`, etc.) that let you create repos, upload files, and call Hugging Face inference APIs (chat, text‑to‑image, translation, streaming, custom endpoints) from browsers, Node, Deno or Bun.

A collection of optimized pre-trained deep learning models and tools for accelerating the development and deployment of high-performance inference applications.

OpenVINO Model Server (OVMS) is a C++ inference server that serves LLMs, vision‑language models, and classic deep‑learning models via OpenAI‑compatible REST or KServe gRPC APIs. Optimized for Intel CPUs, GPUs and NPUs, it provides features like continuous batching, streaming, model versioning, hot‑reload, and Prometheus metrics. Deployable via Docker, bare‑metal binaries, or Kubernetes, it supports many model formats (TensorFlow, ONNX, OpenVINO IR, GGUF, etc.) and integrates with standard client libraries.

A scikit-learn compatible Python library for time series machine learning, providing state-of-the-art algorithms for classification, forecasting, and clustering.

A tool for converting ONNX models into LiteRT, TFLite, TensorFlow, and native PyTorch code to facilitate cross-framework deployment.

A Python package for computations, statistics, and machine learning on manifolds, enabling the analysis of non-Euclidean data like rotations and shapes.

A Python library for time series forecasting that integrates scikit-learn compatible models, statistical methods, and foundation models into a unified API.

A pure-Julia machine learning framework that provides lightweight abstractions for building and training models using native GPU and automatic differentiation support.

A high-level library for PyTorch that simplifies neural network training and evaluation through a flexible engine and event-driven system.

A Python package for probabilistic time series modeling that uses deep learning and PyTorch to generate forecasts with uncertainty intervals.

A fast, header-only C++ machine learning library that provides a wide array of ML methods and functions with bindings for Python, Julia, Go, and R.

A hardware-accelerated JavaScript library for training and deploying machine learning models in the browser and Node.js.

A platform for building AI agents featuring a no-code visual builder and a TypeScript SDK with full two-way synchronization for collaborative agent management.

A multithreaded web crawler that converts website pages into markdown files to simplify document parsing and chunking for RAG and LLM fine-tuning.

An open-source ecosystem for engineered language and structured protocols that provides tools like the Polaris Goal Compiler and Problem Map 3.0 to improve AI execution and debug broken workflows.

A high-performance full-stack self-evolving agent that enables the creation of enterprise AI applications and full-stack apps via natural language and SQL-like service definitions.

An open-source agentic analytics platform that connects LLMs to enterprise data, providing agents with the governed context and tools needed for automated analysis and reporting.

A resilient, HTTP-based data streaming protocol designed for AI agents and collaborative apps to ensure ordered, replayable, and resumable data delivery across devices.

A Python library for building scalable synthetic data pipelines, providing bulk inference and structured output tools to curate high-quality datasets for post-training.

A Ruby gem providing a unified interface for LLMs, vector databases, and prompt management to build RAG applications and AI assistants.

A Python framework for producing structured outputs and building agentic AI workflows using tasks, specialized agents, and type-safe results.

A native-Rust AI agent workspace that provides a unified environment for coding and automation across desktop, terminal, and web interfaces.

An autonomous AI agent orchestrator that manages code workers to solve issues from Linear or local trackers, featuring a Worker/Reviewer pipeline with sandboxed verification.

An enterprise-grade AI UI component library based on Vue 3 and Element-Plus that provides ready-to-use components for chat robots, voice interaction, and streaming interfaces.

DSH Plugin Radar is an open‑source system that automatically discovers, validates (via Kubernetes pods), and catalogs DeepSeek Harness plugins. It produces a live markdown list, curated “featured” lists, bundle collections, and a public JSON API, helping developers and users find reliable DSH plugins without trial‑and‑error.

A .NET wrapper for the PaddleInference C API that enables C# developers to use PaddlePaddle AI models for OCR, object detection, and NLP.

OpenMobius‑skill is a plug‑in for AI coding agents (Claude Code, Codex, OpenClaw, Hermes, Cursor, WorkBuddy) that adds a curated trading‑knowledge base (≈2 000 cards from ICT, SMC, ChanLun, etc.), real‑time market data, 60+ technical indicators, and chart‑generation. It routes user queries through strict/augment/compare modes, supports multiple schools of analysis, and installs self‑contained with its own Python virtual environment, model weights, and Playwright Chromium.

A real-time, browser-based motion capture tool that uses a webcam to drive MMD characters and export animations as .vmd files.

A plug-and-play deployment framework for humanoid robots that decouples controllers, environments, and policies to simplify the deployment of AI control models on real hardware and in simulation.

PlateRecognition is an open‑source, multi‑region licence‑plate detection and OCR system built on YOLO‑v5/v7/v8 and TensorRT. It runs on desktops, servers, Android/iOS and many edge AI chips, delivering >99 % accuracy with sub‑50 ms latency on modern hardware. The project provides C/C++ and Python APIs, supports many languages, and includes a web demo and test code.

PRTS is a Vision-Language-Action foundation model that integrates reward-free contrastive RL into pre-training to give robots a language-grounded sense of goal-reachability.

An open-source vLLM plugin that optimizes LLM and embedding inference to increase throughput and reduce latency using advanced parallelism and speculative decoding.

PhysTwin is a research codebase that reconstructs physics‑informed digital twins of deformable objects from RGB‑D video, provides interactive simulation, analysis tools, and example pipelines for robot planning and batched simulation.

An LLM-powered agent that automates Google Dorking for bug hunting and penetration testing to identify security vulnerabilities and sensitive information disclosures.

A local-first, Git-backed long-term memory system that allows AI coding agents like Claude Code and Codex to share a shared, verifiable Markdown-based memory vault.

A local-first AI workspace for livestream commerce teams that automates creator discovery, livestream transcription, performance review, and video scripting.

A QoS-based scheduling system for Kubernetes that optimizes resource utilization and reliability for hybrid workloads, including latency-sensitive services and batch jobs.

A Neo4j plugin that enables the lossless import, export, and validation of RDF data, bridging the system between W3C semantic web standards and property graphs.