freehire is an open‑source platform that aggregates 3.3 M+ live IT job postings directly from company ATSs, normalises and deduplicates them, and offers faceted search plus AI‑powered career tools (CV builder, fit analysis, chat assistant, auto‑fill extension). Built with Go, PostgreSQL + pgvector, Meilisearch, and a SvelteKit UI, it provides a public key‑less API and a CLI for developers. The stack includes vendor‑agnostic LLM integration, tracing, and observability, making it both a job catalogue and an extensible AI‑assistant for job seekers.

A reference stack for running AI agents safely within NVIDIA OpenShell sandboxes, providing managed inference, network policies, and lifecycle operations.

Moshi is a full-duplex speech-text foundation model designed for real-time, low-latency spoken dialogue, utilizing a specialized streaming audio codec called Mimi.

A simulation framework for evaluating customer service agents across text and voice modalities, testing their ability to follow policies and use tools in real-world domains.

A collection of code and datasets released by Google Research, providing public access to various research implementations.

A set of skills for programming agents to automate the mathematical modeling process, from problem analysis and real code execution to the generation of a formal, verified competition paper.

LiveContext CE is an open‑source, self‑hosted AI automation platform that turns a single chat prompt into a full workflow, scoped LLM agents, data tables and tiny web apps. It bundles a visual workflow engine, AI agents with per‑agent budgets, 700+ built‑in integrations, and optional browser‑rendering add‑ons, all deployable via a single Docker‑Compose stack.

3D Gen Studio is an open‑source, AI‑driven desktop/web app that lets you build 3‑D assets end‑to‑end. It combines a Kanban board or node‑graph UI with ComfyUI workflows and external 3‑D generation APIs, offering image‑to‑mesh pipelines, auto‑UV/retopo, AI‑auto‑rigging (SkinTokens + optional Kimodo motion), LOD/optimization, and game‑ready checks—all stored locally on disk.

An open-source developer platform for Kubernetes that provides high-level abstractions for developers and AI agents to build and deploy apps without managing low-level cluster primitives.

A collection of modular automation skills that extend AI coding agents like Claude Code and Cursor with business capabilities like SEO, market research, and asset creation.

A toolkit that provides AI coding agents with the tools, knowledge, and guardrails needed to build, deploy, and manage applications on AWS.

A self-hosted inference engine that serves multiple open-source models for agent tasks—such as retrieval, OCR, and generation—through a single, unified API.

A collection of human-reviewed instruction sets (skills) for AI agents to ensure they produce idiomatic, secure, and high-performance Go code.

A unified framework to evaluate generative language models on a large number of standard academic benchmarks, serving as the backend for the Open LLM Leaderboard.

HanaAgent is a user-friendly AI assistant with memory and personality that can operate your computer, browse the web, and collaborate across multiple agents via a graphical interface.

An agentic-first RL framework for research that enables training 1T-class MoE models using a PyTorch-native stack with vLLM and Ray.

DarkMoon is an open‑source, self‑hosted platform that uses a fleet of AI agents to run fully automated penetration tests. It orchestrates 50+ real offensive tools via a secure “MCP” gatekeeper, masks all sensitive data before it reaches the language model, and returns reproducible proof for every vulnerability. Supports web, cloud, AD, Kubernetes, databases, IoT and AI/LLM endpoints, runs on local LLMs for privacy, and is GPL‑3.0 licensed.

A lightweight C++20 JIT runtime for NVIDIA CUDA and HUAWEI Ascend NPUs that handles kernel compilation, caching, and loading for device kernel libraries.

An enhanced launcher and management console for the Codex desktop client that provides unified control over model routing, session management, and request logging.

A web search and content extraction suite for the Pi agent that supports dozens of search providers, GitHub repository cloning, and AI-powered video understanding.

NVIDIA Agent Skills is a catalog of portable instruction sets (skills) that teach AI coding agents (Claude Code, Codex, Cursor, etc.) how to install, configure, and use NVIDIA software such as cuDF, cuOpt, DeepStream, Jetson, NeMo, DOCA, and many others. Skills are installed with the open‑source `skills` CLI (`npx skills add nvidia/skills …`) and can be targeted to specific agents. The repo mirrors skill definitions from product repos and is kept up‑to‑date via an automated sync pipeline.

GoRaven is a self-hosted AI agent workstation for teams that enables agents to execute real-world tasks like coding, running commands, and calling APIs instead of just chatting.

A tool and field guide designed to identify and remove generic, AI-generated design patterns (AI slop) from web projects.

A practical node suite for ComfyUI that simplifies MiniMax H3 video generation, offering unified media management and tools for creating long, coherent videos via context segments.

A DeepSeek Harness plugin that provides a comprehensive dashboard and tools for monitoring and managing an AI agent's context window, token usage, and file activity.

An open-source framework for augmenting human productivity by organizing and executing task-specific AI prompts, known as Patterns, via a CLI or REST API.

A Claude Code plugin that queries multiple AI coding agents in parallel to provide side-by-side comparisons and a synthesis of consensus and divergence for technical decisions.

A 3D-printable desktop robot that physically acts out the status of an AI coding agent, providing visual and audio feedback via a REST API.

A local-first control layer for AI coding agents that allows users to manage sessions locally by default with optional multi-device synchronization.

MobileCLIP (and MobileCLIP2) are compact, fast image‑text models for zero‑shot classification and captioning, optimized for mobile devices. The repo offers pretrained checkpoints, training/evaluation scripts (built on OpenCLIP), an iOS demo, and CoCa caption models, all under MIT / Apple research licenses.

ColorTxt (彩读) is a cross‑platform desktop novel reader built with Electron, Vue and Monaco. It reads plain‑text and simple e‑books, adds automatic chapter detection, custom colour‑highlighting, notes, dictionary, translation, multi‑voice TTS, and an AI Reading Assistant. The assistant uses local embedding models (BGE Small ZH, multilingual‑E5 Small) for RAG search and any OpenAI‑compatible chat model for plot analysis, mind‑map/word‑cloud generation, and smart formatting. It can also generate character cards and AI‑drawn portraits via text‑to‑image services. Features include book‑pack export (`.ctz/.ctzx`), WebDAV sync, Pomodoro timer, reading‑ruler, and “fish‑hiding” shortcuts. All AI/ML components can run locally, keeping costs low and preserving privacy.

A real-time interactive streaming engine for digital humans that synchronizes audio and video for lifelike virtual avatars driven by LLMs and TTS.

SAM 3 (Segment Anything with Concepts) is Meta’s 848 M‑parameter foundation model that lets you prompt an image or video with free‑form text (or visual exemplars) and receive masks, boxes, and scores for *all* matching objects. It combines a DETR‑style detector and a SAM 2‑style tracker, introduces a presence token for fine‑grained prompt discrimination, and is trained on >4 M auto‑annotated concepts. The repo provides installation steps, example notebooks, and a new SA‑CO benchmark (270 K concepts) for evaluation.

A tool for bypassing CPU memory protections by scrambling DRAM address translations via the memory controller to expose protected regions like the PSP and SMM.

An open-source framework that enables LLM agents to control Android and iOS devices using natural language, providing tools for UI inspection and mobile automation.

An open-source healthcare AI framework for local clinical text extraction and PII de-identification, supporting multiple platforms including iOS, Android, and web.

An open-source agentic AI data assistant that connects to databases and files to autonomously write SQL and code for data analysis and reporting.

A Chinese‑language, open‑source curriculum that teaches you to build, train, and evaluate large language models from scratch, complete with lecture notes, six hands‑on assignments, and runnable PyTorch code.

A local OSINT research tool for Instagram that analyzes account relationships using a repeated-appearance model to generate ranked candidate lists and network visualizations.

A TypeScript SDK for the Model Context Protocol (MCP), enabling developers to build standardized servers and clients that provide context, tools, and resources to LLMs.

A hook-first Python runtime for building AI agents that live in shared environments like group chats, featuring a transparent tape-based context and pluggable turn pipelines.

A reference implementation for autonomous vulnerability discovery and remediation that uses Claude to find, verify, and patch security bugs in source code.

An AI-native automation layer for EasyEDA Pro that enables AI agents to design schematics and PCBs through a typed action system and a verified circuit block library.

An AI-native writing tool for long-form fiction that uses Agentic RAG and customizable workflows to help authors maintain world-building and character consistency.

A high-performance Go-based proxy that lets you run multiple local AI models and hot-swap between them on demand via a unified API.

Agent Client Protocol (ACP) is a language‑agnostic JSON‑RPC standard that lets code editors talk to generative‑AI coding agents. The repository provides the Rust schema crate, versioned JSON‑Schema files, and links to official SDKs (Kotlin, Java, Python, Rust, TypeScript). It defines a stable protocol version (1) negotiated at runtime, enabling editors and agents to interoperate without custom adapters.

ModSearch is a plug‑in that adds free, real‑time web and X (Twitter) search to LLM chat apps (especially DeepSeek Harness). It uses Firecrawl’s key‑less tier by default (1 000 free credits/month) and can fall back to Antigravity CLI, Tavily, Exa, Grok, or a local fetcher, with automatic key rotation and quota handling. Install with a single `npx` command, then simply ask the model to search or fetch a page; the skill returns a structured JSON response with sources and an uncertainty flag.

Screen Translator is an Android app that captures the screen, runs OCR, translates the text (via on‑device models, LLM APIs, or traditional MT services), and overlays the result back onto the screen. It supports real‑time translation, batch image processing, word‑level dictionary lookup, TTS, and extensive customization, with both offline and cloud options.