Claude Octopus is a Claude Code plugin that adds a multi‑LLM orchestration layer. It stays dormant until you run an `/octo:*` command, then can call up to twelve external providers (Codex, Copilot, Ollama, etc.) to run consensus‑gated reviews, adversarial councils, or a full “spec‑to‑software” pipeline. It ships with dozens of personas, slash commands, and method‑aware workflows, integrates with persistent memory tools, and works in Claude Code, Codex CLI, Cursor IDE (via an MCP server), and OpenCode.

FIM One is an all-in-one agent platform that integrates global SaaS applications with Chinese enterprise stacks to automate complex workflows across disparate systems.

An open-source framework that provides a unified interface for researching and evaluating web agents across multiple benchmarks and open-ended web tasks.

An open-source Obsidian plugin that adds semantic search, an interactive knowledge graph, and an AI agent to help users discover and connect their notes.

A unified TypeScript SDK for building chat bots once and deploying them across multiple platforms like Slack, Discord, and WhatsApp, featuring native AI streaming support.

A Neovim AI coding assistant that integrates LLMs and AI agents into the editor for code generation, refactoring, and review.

A Docker Compose template that bundles n8n, Ollama, and Qdrant to quickly set up a self-hosted, local AI development environment for building AI agents and workflows.

A high-speed GPU document parser that converts images and PDFs into structured Markdown, including tables and formulas, using C++, CUDA, and TensorRT.

A 3D virtual office for AI agents that turns background automation into a visible, immersive workspace for real-time monitoring and collaboration.

An open-source AI poster editor and online design tool that enables users to create professional posters and social media graphics with built-in AI background removal and image generation.

An extension of veRL for training LLM agents via reinforcement learning, featuring a step-independent rollout mechanism to efficiently handle long-horizon, multi-turn tasks.

An open-source, terminal-based coding agent that lets users bring their own local or cloud models for privacy-respecting, local-first AI development.

A lightweight library for easy one-line loading and efficient pre-processing of massive multi-modal datasets for machine learning training and evaluation.

An AI-powered recursive learning system that turns AI coding assistants into interactive tutors using structured knowledge maps and TDD-style exercises.

A JAX-based post-training library for LLMs that provides scalable support for supervised fine-tuning and reinforcement learning on TPUs.

A Model Context Protocol (MCP) server that connects LLMs to ROS-based robots, enabling natural language control and real-time monitoring without modifying robot source code.

GMGN Skills equips AI agents with real-time, multi-chain on-chain data and trading capabilities for meme tokens via CLI commands, enabling automated research, analysis, and execution without scraping.

A development playground and toolkit for the Neuro SAN framework that enables the rapid design and deployment of declarative, multi-agent AI systems.

A knowledge and action layer for Claude Code that enables it to write senior-level .NET 10/C# 14 code using Roslyn-powered analysis and architectural enforcement.

OpenTabs is a CLI‑plus‑Chrome‑extension toolkit that lets LLM agents call real web APIs (Slack, Discord, GitHub, etc.) using the user’s existing browser session, with a library of ready‑made plugins, local‑only execution, and a permission‑based security model.

An industrial deep learning platform providing a unified framework for training and inference, automatic parallelism, and support for scientific computing.

A persistent, shared memory backend for AI agents that provides a knowledge graph and semantic search across sessions and tools via REST API and MCP.

A PyTorch wrapper that abstracts the boilerplate code for distributed training, allowing users to run the same script on any device configuration including multi-GPU and TPU.

An MCP server that integrates professional AI image and video generation into AI coding tools, supporting cloud APIs and local ComfyUI backends.

A modular framework for Agentic Reinforcement Learning that trains multi-step LLM agents by treating each interaction turn as a step-level MDP transition.

An AI-powered tool that analyzes GitHub repositories or local codebases to automatically generate beginner-friendly tutorials and visualizations explaining how the code works.

An efficiency custom nodes package for ComfyUI that integrates and simplifies popular nodes to provide a faster, more streamlined image generation experience.

A framework for organizing project rules and instructions into a verifiable "Skill" to ensure coding agents follow business logic and a strict definition of done.

LoMa is a fast, accurate family of local feature matchers for computer vision — it finds corresponding points between image pairs, serving as a drop-in replacement for LightGlue in 3D reconstruction and visual localization pipelines, with multiple model sizes balancing speed and accuracy.

A serverless, privacy-focused online document editor that allows users to edit Word, Excel, and PowerPoint files locally in the browser using WASM.

A collection of Colab notebooks for fine-tuning and running inference on various open-source LLMs, featuring data prep and advanced training techniques like GRPO.

A zero-dependency TypeScript stock data tool that provides a unified API for A-share, HK, and US stock quotes with native support for AI agents via MCP.

A SvelteKit-based chat interface for LLMs that connects to OpenAI-compatible APIs, featuring smart routing, MCP tool support, and MongoDB persistence.

OWL is a multi-agent collaboration framework for real-world task automation that ranks #1 among open-source frameworks on the GAIA benchmark.

An open-source API layer that turns local LLM inference servers into production-ready AI applications by providing RAG, tool use, and Claude API compatibility.

A library of accountant-reviewed tax guides that AI models can cite to provide verified, professional tax information instead of guessing from training data.

An open-source desktop workspace for personal AI agents that integrates local files, terminals, and a task scheduling system to move beyond simple chat interfaces.

Master‑skill is an open‑source AgentSkill package that lets LLM assistants adopt verified Buddhist master personas. Each master answers queries using real canonical texts (CBETA, BDRC, SuttaCentral, etc.) with citations, enforced by hard‑gate ethics rules and a two‑stage fidelity test suite. Install via `npx master-skill install …`, then invoke slash commands like `/master-huineng` in Claude Code, Cursor, Gemini CLI, or the web chat at fojin.app. The project provides 15 pre‑packaged masters, cross‑tradition comparison tools, offline fallbacks, and a Rust desktop manager.

An end-to-end speech processing toolkit built on PyTorch that provides unified, reproducible recipes for ASR, TTS, speech translation, and audio enhancement.

A Claude/Codex skill that guides an LLM through a four‑step “Mao‑style” analysis of stuck projects, team conflicts, or life decisions, outputting a concise text summary and an optional single‑file HTML report.

Open Storyboard Canvas is an open‑source, cross‑platform desktop app (Tauri + React + Rust) that provides a visual node canvas for AI‑generated images, videos, and 3‑D storyboard scenes. It includes a chat‑based Canvas Agent (beta) that can create, edit, and track generation tasks, supports multiple providers (OpenAI‑compatible, Claude, Dreamina CLI), and offers director‑studio tools, prompt libraries, bulk import, and detailed logs. Licensed MIT, built on the Storyboard‑Copilot upstream project.

FlashMLA is DeepSeek’s CUDA library of dense and token‑level sparse attention kernels (including FP8 KV‑cache support) optimized for NVIDIA SM90/SM100 GPUs. It powers DeepSeek‑V3/V3.2/V4 models, delivering up to 660 TFLOPs on H800 GPUs and offering a fused kernel that combines Q‑norm, RoPE, attention and FP8 casting. Installation is via a standard pip build, and usage follows simple Python APIs (`flash_mla_with_kvcache`, `flash_mla_sparse_fwd`, etc.).

An AI API proxy and management platform that provides unified access, load balancing, and cost tracking for multiple AI providers like Claude and OpenAI.

LLM4AD Next is a Python toolkit that turns a natural‑language problem description into a full, runnable evolutionary‑algorithm search project. An interactive LLM consultant generates the code skeleton, configuration, and evaluator, then built‑in evolutionary methods (IslandGA, Diverse Island GA, EoH, etc.) automatically evolve the generated algorithm. It also offers an advisor to suggest evolution‑ready code blocks, long‑term memory support, Docker images, and an online demo.

A CLI tool that allows users to customize system prompts, UI themes, and functional behavior of Claude Code by patching its installation.

An autonomous research agent that uses a hypothesis-tree framework to optimize software and ML tasks through disciplined, automated experimentation and validation.

MCPJam Inspector is an open‑source testing and evaluation platform for servers that implement the Model‑Chat‑Protocol (MCP) used by chat‑based AI products. It provides a web playground, OAuth debugger, server‑debug tools, reusable “skills”, shared workspaces, eval test‑case runner, a CLI, an SDK, and CI/CD integrations to catch regressions across 16 client configurations and 170+ LLM models.

BCR Bot is a simulation package for a differential‑drive robot with IMU, depth, stereo cameras and 2‑D LiDAR, supporting ROS 1/2 across multiple Gazebo releases and NVIDIA Isaac Sim, plus ready‑made mapping and navigation examples.