An AI Agent system for long-form story creation and translation that manages world-building, character continuity, and narrative state to prevent plot holes and AI-style prose.

Codex Router is a community‑maintained bridge that lets the Codex editor/CLI send prompts to many external LLM providers (Anthropic, Kimi, DeepSeek, Grok, Gemini, Claude, etc.). It stores credentials locally, offers a guided installer, a macOS menu‑bar/desktop‑widget UI, and a Homebrew‑installable CLI. Experimental bridges let it reuse existing Claude, Cursor, and Gemini agents without exposing their OAuth tokens.

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

An extremely fast Python package and project manager written in Rust that replaces pip, poetry, and pyenv to streamline dependency and version management.

A collection of AI agent skills for Codex and Claude Code that automates the end-to-end production of polished Three.js browser games, from asset generation to QA.

An open-source, local-first AI video editor that combines a professional multitrack timeline with conversational agents for editable, agent-driven video production.

Tick‑Stock‑Panel (TSP) is a self‑hosted, Docker‑deployable web platform for Chinese A‑share quantitative research. It pulls data from multiple plug‑in sources, routes capabilities, computes 68 technical indicators with Polars, runs 25+ built‑in (and custom/AI‑generated) stock‑selection strategies, provides factor editing, mining, and back‑testing (including minute‑level simulation), and offers real‑time monitoring with voice/Feishu alerts. The stack uses FastAPI, DuckDB/Parquet, Polars, and a React 18 front‑end; optional AI integration works via any OpenAI‑compatible API. The README details quick‑start Docker commands, dev mode, and configuration via a `.env` file.

A single-file, dependency-free reverse proxy that exposes Command Code's proprietary API as OpenAI- and Anthropic-compatible endpoints, with streaming, tool calling, multimodal image input, and anti-detection device-fingerprint emulation, built by analyzing the official CLI's traffic.

LingBot‑World‑V2 is a research‑grade video‑generation model that creates infinite, interactive worlds from text prompts. It offers fast (real‑time) and high‑quality variants (14 B and 1.3 B), supports diverse actions, and uses a pilot‑director agent architecture. The repo provides inference code, model download links, and a quick‑start guide; deployment code is not released.

A prompt-based toolkit that turns AI assistants into learning companions by forcing users to think and reason through code rather than just generating solutions.

OpenChamber is an open‑source, cross‑platform workspace that lets developers start AI coding agents (via OpenCode), monitor their progress, compare multiple model runs, review changes with guided walkthroughs, preview the app being edited, and integrate directly with GitHub—all while syncing sessions across desktop, web/PWA, VS Code, and mobile devices.

An efficient closed-loop harness for embodied AI that evolves code-based recovery skills and critics to improve robot task success rates without retraining the base policy.

A cross-platform desktop client that provides a unified interface for multiple LLM providers, including cloud services, web AI, and local models.

EnvoyMesh is a self‑hosted, peer‑to‑peer social network that lets you chat, call, share files, and run a personal AI assistant without any central server. Your identity is a cryptographic Ed25519 key you control, and the built‑in AI (OpenClaw) or any external HTTP‑based AI can act on your behalf under strict policy checks. The platform includes a markdown knowledge base (with Obsidian sync), a mobile thin client (EnvoyGo), and team‑oriented agent orchestration features.

A collection of AI agent skills for transforming novels into short dramas, automating the pipeline from outline and character design to script and storyboard.

An AI‑agent skill that lets coding assistants review and improve UI designs against Apple’s Human Interface Guidelines, supporting Flutter, React Native, SwiftUI, etc., and providing framework‑specific fixes.

An AI-powered penetration testing engineer for authorized security assessments that automates the discovery of business-logic flaws and cloud misconfigurations across global and Chinese cloud providers.

LibreTranslate is a self-hosted, open-source machine translation API that removes dependency on proprietary services like Google or Azure.

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.

An open-source AI chatbot framework for ESP32 microcontrollers that enables voice interaction with LLMs and hardware control via the MCP protocol.

A framework for distilling an author's writing style into a structured "Writing DNA" that AI agents can use to stably and accurately replicate that style.

ABot-Recon is a streaming 3D reconstruction system that uses a fixed 12-frame local context to efficiently build global trajectories and point clouds from long video streams without needing persistent long-range memory.

PentestGPT is an MIT‑licensed Python CLI that chains LLMs (Claude Code, Codex, or many other providers) into an autonomous penetration‑testing pipeline. It runs staged phases (recon → exploit → walkthrough for CTF, or asset discovery → vulnerability identification → report for real pentests), can persist sessions, and offers a legacy interactive mode where a user guides three cooperating LLMs. Install with `make install` (Python 3.12+, uv) or via Docker, then run `pentestgpt --target <IP>` (or `pentestgpt-legacy` for the interactive version). The tool reports optional anonymous telemetry and achieved ~86 % success on a published benchmark.

An intelligence runtime and infrastructure layer for AI agents that provides collective error recovery, persistent memory, and self-sovereign cryptographic identities.

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.

An AI-driven short-drama production workflow consisting of eleven agent skills that transform ideas or novels into scripts, storyboards, and video prompts with a focus on visual consistency.

Pipecat is an open‑source Python framework for building real‑time voice and multimodal AI agents. It offers plug‑and‑play support for dozens of speech‑to‑text, large‑language‑model, text‑to‑speech, and speech‑to‑speech services, composable pipelines, multi‑agent orchestration, and client SDKs for web, mobile, and embedded platforms. The Pipecat CLI can scaffold a working bot in minutes, and additional tools (Whisker debugger, Tail dashboard, Voice UI Kit) help with development and deployment.

A set of AI agent skills that enable end-to-end hotel search, live rate comparison, and booking management across global inventory.

Kiro Crew is an open‑source, on‑premise AI‑agent platform that keeps agent state (memory, lessons, skills) alive across restarts, lets you run long‑running or scheduled tasks unattended, and is reachable via a desktop app, web dashboard, CLI, or chat integrations (Slack, Discord, etc.). Install with a one‑line script, Docker, or the native desktop packages; the Gateway process stores all data locally and enforces sandboxing and approval policies.

cmux is a native macOS terminal (built on Ghostty) that adds vertical tabs, a notification system, an in‑app scriptable browser, and SSH workspaces to help developers run and monitor multiple AI coding agents (Claude Code, Codex, OpenCode, etc.) in parallel.

An open-source voice-to-text dictation app for macOS that enables on-device AI transcription and system control via voice commands.

An AI-assisted reverse engineering toolkit that prevents hallucinations by verifying AI-proposed hypotheses against the actual bytes of a binary.

A high-performance computer vision framework implementing YOLO models for real-time object detection, segmentation, classification, and pose estimation.

RL training environments for the Microduck bipedal robot that use high-fidelity actuator modeling and domain randomization to enable sim2real transfer of complex gaits and tricks.

A native desktop app that aggregates coding-agent sessions from multiple tools into a single, searchable, and resumable local interface.

An open-source AI-powered social media scheduling tool that allows users to manage, automate, and analyze posts across multiple platforms.

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

An agent skill that converts complex terminal output, such as diagrams and tables, into styled, interactive HTML pages for better readability.

A local-first personal context workspace that organizes files, conversations, and memory into scoped "Context Rooms" to provide AI agents with grounded, traceable context.

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

A read-only, in-database observability tool for PostgreSQL that provides health reports and performance diagnostics without requiring an agent or external service.

An open-source AI Agent platform for Android that integrates LLMs with system automation, a built-in Ubuntu terminal, and project workspaces for mobile productivity.

A 3D desktop pet driven by a live spiking simulation of the real fruit fly connectome, reacting to cursor movements and window layouts using biological neural circuits.

A Windows application that applies NVIDIA DLSS 5 Neural Rendering and Frame Generation to images and videos via a local Gradio interface.

A framework for programming rather than prompting language models, allowing developers to build modular AI systems with automatically optimized prompts and weights.

A framework that decomposes single anime illustrations into up to 23 semantically distinct, inpainted layers exported as a PSD for 2.5D animation and rigging.

LingBot‑Map is a feed‑forward 3‑D reconstruction model that streams video frames through a Geometric Context Transformer, using a paged KV‑cache (FlashInfer) to keep memory low. It runs at ~20 fps on modest GPUs, supports very long sequences via keyframe‑interval or windowed inference, and ships with ready‑to‑run demos, an offline batch renderer, and pretrained checkpoints on HuggingFace/ModelScope.

A persistent memory layer for AI agents that stores explicit facts with sources and uses a knowledge graph to provide synthesized answers instead of simple search results.