A tool for generating multi-view optical illusions using pixel-based diffusion models, creating images that change identity when rotated, flipped, or rearranged.

A high-performance Rust implementation of GraphRAG that enables building knowledge graphs from documents for natural language querying across server, WASM, and hybrid deployments.

A local RAG chatbot that uses llama.cpp and Chroma to answer questions based on a Markdown-based knowledge base with incremental indexing.

HiRAG is a Retrieval-Augmented Generation framework that uses hierarchical knowledge to improve the quality and comprehensiveness of LLM responses compared to standard RAG.

A lightweight and modular Python library for implementing Retrieval-Augmented Generation (RAG), allowing users to easily integrate various LLMs and vector stores to chat with their documents.

A cross-platform computer vision framework for tracking fiducial markers and multi-touch fingers to enable the creation of tangible user interfaces.

The official Python package for Roboflow, enabling developers to interact with models, datasets, and projects to build and deploy computer vision models programmatically.

A real-time computer vision library written in Java that provides tools for low-level image processing, camera calibration, and feature tracking.

A PHP library for Generative AI and vector databases, providing a unified interface to integrate LLMs and vector stores into PHP applications.

Hallo-Live is a real-time text-driven framework that generates synchronized audio and video for digital avatars using a causal dual-stream Diffusion Transformer.

FoleyCrafter is a video-to-audio generation framework that produces realistic, synchronized sound effects based on the visual content of a silent video.

A plug-and-play framework for Diffusion Transformers that enables the generation of longer videos and higher-resolution images through positional embedding extrapolation.

A training-free iterative framework for long story visualization that maintains semantic consistency across up to 100 frames using a global reference cross-attention module.

MACE is a finetuning framework for diffusion models that enables the mass erasure of up to 100 unwanted concepts, such as celebrities or explicit content, while preserving unrelated model knowledge.

A library for improving LLM performance in Retrieval Augmented Generation (RAG) by creating augmented datasets and fine-tuning models using PEFT.

An AI datastore for schemas, BLOBs, and predictions that provides tools for human supervision and multi-modal data labeling.

A TensorFlow-based handwritten text recognition system that converts images of single words or text lines into digital text using CNN and LSTM layers.

An AI-powered web crawler that uses natural language prompts to discover relevant pages and extract structured JSON or Markdown data from websites.

An LLMOps template for Prompt flow that streamlines the end-to-end lifecycle of LLM applications, from local experimentation and evaluation to production deployment.

An LLMOps framework for building AI applications that provides tools for prompt templating, chaining, and agentic tool integration.

An open-source observability library for JavaScript/TypeScript LLM applications that uses OpenTelemetry to trace calls to LLM providers and vector databases.

A tracking system and dashboard that compares the success rates of AI coding agents by analyzing their pull request merge rates on GitHub.

RayDP allows running Apache Spark on Ray, enabling the creation of unified, distributed data and AI pipelines in a single Python program.

A suite of AI-powered tools that automatically enhance Anki flashcards with AI-generated images, mnemonics, and explanations to improve learning efficiency.

HistomicsTK is an open‑source Python toolkit for analyzing whole‑slide pathology images. It offers stain normalization, nuclei segmentation, and feature extraction, usable as a standalone library or as a plugin for the Digital Slide Archive web platform.

NagaAgent is a cross‑platform desktop AI assistant (Windows/macOS/Linux) that offers chat with any OpenAI‑compatible LLM, custom tool‑calling via JSON blocks, graph‑RAG memory stored in Neo4j, a Live2D animated avatar, voice synthesis/recognition with lip‑sync, a plug‑in MCP framework for weather, web search, game guides, etc., and community features like a forum and skill marketplace. It ships as an Electron/Vue front‑end plus Python FastAPI back‑end, is installable via `uv sync` (or pip), and is dual‑licensed (AGPL‑3.0 + proprietary).

ResShift is an efficient diffusion model for image super-resolution and restoration that reduces sampling steps to improve inference speed without sacrificing image quality.

An agentic operating system for Snowflake Coco that brings structured, multi-agent workflows and governance-aware execution to data engineering projects.

A framework for building and running reusable, typed AI procedures called "methods" using a declarative .mthds file format to orchestrate complex AI workflows.

An open-source ROS-based framework and hardware kit that provides tools and preconfigured images to easily create and program autonomous drones.

Unbody is an open-source backend infrastructure project designed to provide a comprehensive data management and indexing system for AI applications.

An open-source extension of the AWS CDK that provides well-architected infrastructure patterns for quickly deploying generative AI solutions on AWS.

A CLI tool for scaffolding production-ready LLM training projects, providing templates for different model sizes and a complete pipeline from tokenization to deployment.

A JavaScript-based reference implementation of a RAG application that combines Azure OpenAI and Azure AI Search to create a ChatGPT-like experience over enterprise data.

A diffusion model for generating novel protein backbone structures, enabling the creation of new protein folds for computational protein design.

A high-performance inference engine that uses SVDQuant to enable 4-bit quantization for diffusion models, significantly reducing VRAM usage and increasing inference speed.

A set of AI skills and templates that guide AI agents to generate high-quality, professional personal homepages, portfolios, and HTML presentations instead of generic templates.

An AI-powered English speaking assistant based on Google Gemini that provides real-time pronunciation feedback and grammar corrections.

A text-based game generator and sandbox environment used for training and testing reinforcement learning agents in simulated text worlds.

A deterministic output compactor for terminal-heavy AI agent workflows that reduces noisy command output to save tokens and improve context clarity.

A low-cost, universal leader-follower teleoperation system that allows users to control a wide variety of commercial robot arms using a 3D-printed leader arm.

A Python library implementing Natural Gradient Boosting for probabilistic prediction, allowing models to output probability distributions instead of single point estimates.

A financial market simulation engine powered by a generative foundation model that simulates individual order events to create realistic emergent market behavior.

An MCP server that integrates Google Calendar with AI assistants, enabling them to manage schedules, detect cross-account conflicts, and create events from images or PDFs.

An orchestration framework that runs multiple Claude Code agents in parallel to systematically fix bugs and implement best practices across large codebases.

A framework for building MCP servers that automatically converts Python files into server capabilities, reducing boilerplate and providing built-in authentication and telemetry.

An AI-powered IDA Pro plugin that assists binary reverse engineering through LLM-driven function explanation, semantic knowledge graphs, and RAG-based document search.

YunChang provides a Unified Sequence Parallel (USP) attention mechanism that hybrids DeepSpeed-Ulysses and Ring-Attention to enable efficient long-context LLM training and inference.