A native Zig-based inference server for Apple Silicon that runs LLMs and multimodal models with OpenAI, Anthropic, and Ollama API compatibility.

A Rust-based vector index with Python bindings that uses the TurboQuant algorithm to provide high-compression, fast SIMD search without requiring a separate training phase.

A set of agent skills and SDKs for driving a modified Chromium browser that mimics real device fingerprints to bypass bot detection and manage isolated profiles.

An open-source, local-first workspace that coordinates multiple AI coding agents through shared channels, task boards, and long-term memory.

The official Lark/Feishu CLI tool designed for humans and AI Agents to interact with 18+ business domains and 2,500+ APIs through a structured command system.

A high-performance vector database and similarity search engine written in Rust, designed for storing and retrieving embeddings with advanced filtering and hybrid search capabilities.

Switchyard is a pre‑1.0 NVIDIA‑NeMo project that routes each LLM request to the cheapest model that can still solve the task. It offers a Rust library, a Python package, a NeMo Relay plugin, and a demo OpenAI‑compatible proxy. Configurable routing algorithms (stage, capability, escalation, etc.) let you automatically fall back from cheap to strong models, cutting token cost by up to ~30 % while keeping accuracy near that of a single high‑end model. The code is Apache‑2.0 licensed and intended for integration into existing LLM gateways or custom harnesses.

A Memory Agent that manages the long-term memory of AI agent fleets by curating, reconciling, and expiring information to prevent data noise and vendor lock-in.

An AI-powered data visualization tool that uses agents to connect fragmented data sources and a branching "Data Thread" workspace for natural language data exploration.

An enterprise-grade MCP and A2A gateway that provides secure, governed access for AI copilots and autonomous agents to internal enterprise tools and data.

A peer-to-peer LLM inference system that splits large models across multiple browser tabs on different devices, enabling the execution of massive models without central servers or installations.

A high-performance, GPU-accelerated terminal workbench that provides persistent sessions and deep integration for AI coding agents.

A collection of reference implementations for the Model Context Protocol (MCP), providing examples of how to give LLMs secure access to tools and data sources.

An AI-guided toolkit for authorized security testing that provides a curated library of 215 tools and translates plain English into specific commands and execution plans.

A TypeScript library for WebGPU that provides typed shader imports and a unified API for running GPU code in the browser, headless Node, and test environments.

A local AI server stack that provides a one-command installation of inference, chat, RAG, and agent tools for private AI on your own hardware.

A security evaluation framework for generative AI that uses adversarial testing and runtime evidence to detect security boundary crossings in LLMs, RAG apps, and AI agents.

A production-ready headless browser built in Rust for AI agents that uses on-demand rendering to drastically reduce resource consumption during web automation and extraction.

A tool that converts natural language, code, and infrastructure configs into editable .drawio diagrams, featuring incremental synchronization and architecture-as-test validation.

A graphical desktop management tool for CLIProxyAPI that aggregates multiple AI API providers and simplifies agent client configuration.

An open-source LLM observability and evaluation platform for AI agent tracing, prompt management, and production monitoring.

A commercialization desktop client based on DeepSeek Harness and Jingyun Studio that enables AI developers to turn agents and workflows into monetizable products.

An on-device hybrid search engine that indexes local markdown notes and documentation using BM25, vector search, and LLM reranking for high-precision retrieval.

User Scanner is a Python CLI/Library that enumerates email addresses and usernames across 1 080+ online services, extracts rich profile metadata, pivots discovered handles, checks breach logs, and can be driven by LLM agents through a Model Context Protocol server.

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.

Graphiti is a framework for building temporal knowledge graphs that track how facts change over time, enabling AI agents to maintain evolving, high-provenance context.

A local-first search layer that unifies ripgrep, BM25, and vector search to provide high-precision retrieval for both humans and AI agents.

An open-source integration layer that centralizes tool configurations, authentication, and policies for AI agents via the Model Context Protocol (MCP).

An evidence-gated analytics system that ensures rigorous data validation and reproducible evidence are established before generating decision intelligence reports.

A local inference router that distributes AI workloads across multiple compatible computers on the same network using OpenAI and Ollama compatible endpoints.

CloakBrowser is a custom‑patched Chromium binary that looks like a normal Chrome to anti‑bot services. It ships with Python, Node, and .NET wrappers that act as drop‑in replacements for Playwright/Puppeteer, letting you run web‑automation (including AI agents) without being flagged. The binary is auto‑downloaded, verified, and includes 87 source‑level stealth patches plus a `humanize` mode for realistic mouse/keyboard actions. Free for one concurrent session; Pro licenses enable scaling.

A tool that unlocks DLSS 3 Frame Generation and Neural Rendering for NVIDIA RTX 20 and 30 series GPUs in DirectX 12 games.

A high-performance AI gateway that unifies 23+ AI providers under a single OpenAI-compatible API with automatic failover and semantic caching.

A context engine that indexes codebases into knowledge graphs to provide AI agents with deep architectural awareness and impact analysis tools.

A high-performance Rust-based statusline HUD for Claude Code that displays real-time model info, reasoning effort, Git status, and context window usage.

An infrastructure layer and unified API that turns existing agent harnesses into plug-and-play backends, allowing products to switch between different AI agents without custom engineering.

A Model Context Protocol (MCP) server that connects AI agents and chatbots to GitHub, enabling them to manage repositories, issues, and pull requests via natural language.

A portable toolkit and agentic framework for automating the discovery, reproduction, and patching of software vulnerabilities to build more secure code.

A fast Rust library for PDF classification and text extraction that converts PDFs to clean Markdown and uses smart routing to avoid unnecessary OCR.

A tiny, agent-native CLI for generating text, images, video, and audio using the Vercel AI SDK and AI Gateway for unified model access.

A workflow and runtime layer for OpenAI Codex CLI that adds structured planning, durable state tracking, and specialized roles for complex coding tasks.

A full application framework for the Model Context Protocol (MCP) that simplifies building servers, clients, and interactive apps to connect LLMs to tools and data.

An AI-powered OSINT username scanner that uses machine learning and heuristics to find profiles across 840+ platforms with higher accuracy than traditional tools.

A persistent, file-based planning system for AI coding agents that stores task plans and progress on disk to prevent memory loss during context resets.

An autonomous AI penetration testing platform that uses an LLM agent and an independent verifier to discover and prove vulnerabilities with concrete evidence.

Agent Substrate is a Kubernetes‑native runtime that multiplexes thousands of stateful AI agents onto a few sandboxed worker pods, offering sub‑500 ms resume times, full memory/filesystem snapshotting, and zero‑trust isolation via gVisor or micro‑VMs. It’s framework‑agnostic, provides a CLI (`kubectl‑ate`), and includes demos and integrations (Agent Executor, kagent). The project is early‑stage and not yet production‑ready.

An MCP server that integrates IDA Pro with LLMs, allowing AI assistants to perform reverse engineering tasks like decompilation, renaming, and binary analysis.

An open-source AI agent for SREs that automates production incident investigation and root cause analysis across any infrastructure stack.