AReaL: a scalable asynchronous RL infrastructure for training large-scale reasoning and agentic models
A large-scale asynchronous reinforcement learning system designed for training high-performance reasoning and agentic AI models.
jcode: a high-performance coding agent harness with semantic memory and collaborative multi-agent swarms
A high-performance coding agent harness featuring semantic memory, multi-agent swarms for collaboration, and an extremely resource-efficient CLI interface.
Open-Interface: a cross-platform computer automation tool that uses LLMs to simulate mouse and keyboard inputs
An AI-powered computer automation tool that uses LLMs to self-drive a PC by simulating keyboard and mouse inputs based on visual feedback.
terrain-diffusion: a learned successor to Perlin noise for infinite and deterministic real-time terrain generation
A learned successor to Perlin noise that uses diffusion models to generate infinite, deterministic, and randomly-accessible real-time terrain and climate data.
SkillOpt: an executive strategy for self-evolving agent skills that optimizes skill documents using deep-learning optimization principles
SkillOpt is a framework that treats agent skill documents as trainable states, using an optimizer model to evolve them through a rigorous training loop without modifying model weights.
BallonsTranslator: a deep learning-assisted comic translation tool with automated inpainting and typesetting
A deep learning-assisted comic translation tool that automates text detection, image inpainting, translation, and typesetting for manga and comics.
ccg-workflow: a multi-model orchestration engine for Claude Code that automates complex coding workflows with persistent state and quality gates
A workflow engine for Claude Code that turns it into a multi-model orchestrator, dispatching specialized tasks to Codex, Gemini, and Grok for parallel analysis and implementation.
dramaclaw: an industrialized AI production line for transforming manuscripts into finished short dramas
An industrialized AI production pipeline that transforms manuscripts into finished short dramas by automating script generation, storyboarding, voice-over, and video composition.
fulling: a foundation for persistent AI workspaces with secure Kubernetes credential boundaries
Fulling is a foundation for building dedicated AI workspaces that combine runtime, memory, and skills within a secure Kubernetes-based credential boundary.
trainer: a Kubernetes-native distributed AI platform for scalable LLM training and fine-tuning
A Kubernetes-native distributed AI platform for scalable LLM fine-tuning and training across various frameworks like PyTorch, JAX, and XGBoost.
maxtext: a high-performance JAX-based reference implementation for scalable LLM pre-training and post-training
A high-performance, scalable LLM library and reference implementation written in JAX for training and post-training models on TPUs and GPUs.
oumi: an end-to-end open-source platform for training, evaluating, and deploying foundation models
An open-source platform that streamlines the end-to-end lifecycle of foundation models, providing tools for data curation, training, evaluation, and deployment.
openvino: an open-source toolkit for optimizing and deploying deep learning models across diverse hardware platforms
An open-source toolkit for optimizing and deploying deep learning models across CPU, GPU, and NPU hardware to boost inference performance.
aimet: AI Model Efficiency Toolkit – Quantization & Compression for Edge Deep Learning
AIMET is Qualcomm’s open‑source toolkit for quantizing and compressing PyTorch/ONNX models, enabling fast, low‑memory inference on edge devices with minimal accuracy loss.
TransformerEngine: a library for accelerating Transformer models on NVIDIA GPUs using low-precision numerical formats
A library for accelerating Transformer models on NVIDIA GPUs by utilizing low-precision formats like FP8, MXFP8, and NVFP4 to improve performance and reduce memory usage.
huggingface_hub: the official CLI and Python client for interacting with the Hugging Face Hub platform
The official Python client and CLI for the Hugging Face Hub, enabling users to programmatically download, upload, and manage ML models, datasets, and demos.
lance: an open lakehouse format for multimodal AI with high-performance vector search and random access
An open lakehouse format for multimodal AI that provides high-performance vector search, random access, and native storage for images, video, and audio.
onnx: an open ecosystem and format for AI model interoperability across frameworks and hardware
An open source format for AI models that enables interoperability between different frameworks and tools, streamlining the path from research to production.
nono: a zero-latency agent sandbox that isolates both the agent and the tools it calls
A zero-latency, zero-setup sandbox for running AI agents securely with least-privilege access and isolated tool execution.
cli-printing-press: an automated factory that generates agent-optimized CLIs and MCP servers with local SQLite persistence
A factory for generating token-efficient Go CLIs and MCP servers for any API or website, featuring local SQLite persistence for complex agent-native queries.
agent-scripts: a centralized hub for managing shared AI agent rules and reusable workflow skills across local workspaces
A centralized management system for shared AI agent instructions, reusable skills, and portable helper scripts across local workspaces.
OpenAgentsControl: a context-aware AI coding framework that enforces team coding patterns through human-guided approval gates
An AI coding framework that teaches agents your specific coding patterns to generate production-ready code that requires no refactoring, featuring human-guided approval gates and token-efficient context management.
nixtla: a foundation model for zero-shot time series forecasting and anomaly detection
A foundation model for time series forecasting and anomaly detection that provides zero-shot inference and fine-tuning capabilities across multiple domains.
infinity: an AI-native database providing high-performance hybrid search of dense and sparse embeddings, tensors, and full-text
An AI-native database that provides high-performance hybrid search across dense and sparse embeddings, tensors, and full-text data for LLM and RAG applications.
PixelRAG: what it is, what problem it solves & why it's gaining traction
PixelRAG is a visual RAG system that renders documents as screenshots instead of parsing them to text, allowing AI to retrieve and reason over visual elements like tables and charts.
nexent: a zero-code platform for auto-generating production-grade AI agents using natural language
A zero-code platform for auto-generating production-grade AI agents using natural language prompts, featuring multi-agent collaboration and layered memory.
cactus: a hybrid edge-cloud AI engine for mobile devices and wearables with custom rotation-based quantization
A hybrid edge-cloud AI engine for mobile devices and wearables that provides high-performance local inference for text, speech, and vision with automatic cloud handoff.
DeepTutor: what it is, what problem it solves & why it's gaining traction
DeepTutor is an agent-native personalized tutoring workspace that integrates tutoring, research, and mastery practice into a single system with shared memory and multi-engine RAG.
opendataloader-pdf: a high-accuracy PDF parser for AI-ready data extraction and automated PDF accessibility tagging
A high-accuracy PDF parser and accessibility tool that converts PDFs into structured Markdown, JSON, or Tagged PDFs for AI pipelines and screen-reader compliance.
ruby_llm: a unified Ruby AI framework for building multimodal chats, agents, and agentic workflows in Rails
A unified Ruby AI framework that provides a consistent interface for multiple LLM providers, enabling the creation of chats, agents, and multimodal AI features in Rails applications.
openagents: a collaborative workspace and launcher for coordinating multiple AI agents in a single shared environment
A collaborative OS for AI agents that provides a unified workspace where different AI agents can collaborate, share files, and share a live browser.
GLM-5: a series of flagship LLMs optimized for long-horizon agentic engineering and complex systems tasks
A series of flagship open-source LLMs optimized for long-horizon agentic tasks, complex systems engineering, and coding, featuring a 1M-token context window in GLM-5.2.
obsidian-skills: what it is, what problem it solves & why it's gaining traction
A collection of agent skills that enable AI agents like Claude Code and Codex to create and edit content within Obsidian vaults using Obsidian-specific syntax and the JSON Canvas format.
awesome-llm-apps: what it is, what problem it solves & why it's gaining traction
A comprehensive cookbook of 100+ ready-to-run AI agent and RAG templates that developers can clone and customize to ship production LLM apps.
minds: a unified workspace to delegate projects to AI agents and publish their outputs as web apps
A unified workspace for delegating complex projects to AI agents and converting their outputs into publishable web applications and dashboards.
manga-ocr: a Japanese OCR engine optimized for manga text bubbles and complex layouts
An OCR tool specifically optimized for Japanese manga, capable of recognizing multi-line text, furigana, and both vertical and horizontal layouts.
ML: a high-level machine learning and deep learning library for the PHP language
A high-level machine learning and deep learning library for PHP that provides over 40 algorithms and tools for the entire ML lifecycle.
voice-pro: an all-in-one AI dubbing studio for YouTube processing, voice cloning, and multilingual translation
An AI-powered web application for speech recognition, translation, and multilingual dubbing that integrates ASR, TTS, and voice cloning into a single workflow.
ai-job-search: an AI-powered job application framework that automates tailoring, PDF layout verification, and ATS-compatibility checks for resumes
An AI-powered job application framework that automates job searching, tailoring LaTeX CVs and cover letters, and interview preparation using a drafter-reviewer workflow.
OM1: a modular AI runtime for deploying multimodal agents across physical robots and simulators
A modular AI runtime for creating and deploying multimodal AI agents across digital environments and physical robots, including humanoids and quadrupeds.
voltagent: an end-to-end AI agent engineering platform with a TypeScript framework and production observability console
An end-to-end AI agent engineering platform featuring a TypeScript framework for building agents and a console for observability, deployment, and operations.
unity-mcp: an MCP-based bridge that enables LLMs to control the Unity Editor via natural language
A bridge that connects AI assistants to the Unity Editor via the Model Context Protocol, enabling developers to manage scenes, assets, and scripts using natural language.
MemOS: a memory operating system for AI agents providing persistent multi-modal storage and graph-based memory management
A memory operating system for LLMs and AI agents that provides persistent, multi-modal long-term memory and the ability to evolve skills through experience.
meetily: what it is, what problem it solves & why it's gaining traction
Meetily is a privacy-first, open-source AI meeting assistant that provides local transcription and AI-powered summaries to ensure sensitive meeting data never leaves the user's device.
Vibe-Trading: a personal trading agent framework with quantitative factor libraries and multi-broker execution capabilities
An AI-powered trading agent framework that integrates quantitative factors, fundamental data, and broker connectors to automate research and trade execution.
E2B: secure isolated cloud sandboxes for executing AI-generated code
An open-source infrastructure for running AI-generated code in secure, isolated cloud sandboxes via Python and JavaScript SDKs.
9router: a smart AI router with auto-fallback and token compression to minimize costs and avoid rate limits
A smart AI router and token saver that provides auto-fallback between subscription, cheap, and free AI models to prevent rate limits and reduce costs.
llmfit: a hardware-aware model recommender that right-sizes LLMs to your system RAM and GPU
A terminal tool that detects your hardware and recommends the best-fitting LLM models based on memory, speed, quality, and context.
nanobot: a self-hosted personal AI agent runtime with multi-channel chat integration and persistent memory
An ultra-lightweight, self-hosted personal AI agent runtime that provides a persistent agent with memory, tools, and integrations across multiple chat platforms.
JeecgBoot: an AI-driven low-code platform that generates enterprise Java systems from natural language prompts
An AI-powered low-code development platform for Java enterprises that enables system generation via natural language and combines zero-code configuration with automated code generation.