Explore the directory

A lightweight Python framework for building LLM-based agents, modeled on neural-network-style layers where agents pass typed messages and automatically record them to memory.

• Agent frameworks

An agent that manages the shared memory of a fleet of AI agents: extracting durable facts from their conversations, merging duplicates, resolving contradictions, forgetting stale information, and briefing each agent before it acts.

• Agent memory

A portable memory-and-skills folder that keeps a coding agent's context and conventions consistent across more than a dozen harnesses, plus bounded, budgeted agentic loops and a local dashboard for agent activity and cost.

• Agent skills
• Agent memory

A local memory and token-accounting layer that keeps task context, a project map, and known fixes available across Claude Code, Codex and OpenCode sessions, without calling any external API.

• Agent memory

An MCP server that lets Claude and other AI assistants call out to Google's Gemini CLI, or its successor Antigravity CLI, to analyze large files and codebases with a bigger context window.

• MCP servers

A set of eleven Claude Code and Codex skills that turn a novel or idea into a short serialized drama or comic: episode scripts, visual style guides, storyboards, and image or video generation prompts, with confirmation before any paid step.

• Agent skills
• Generative media

A structured, evaluation-backed workflow of skills for AI coding agents that classify a request, define what done means, gather evidence, and verify results before reporting, distilled from an internal Claude workflow.

• Agent skills

A Chinese-language tutorial series that teaches how to build large language models from scratch, covering pretraining, fine-tuning, mixture-of-experts, RLHF and deployment, with accompanying notebooks and videos.

• Tutorials

A hook and skill guard for AI coding agents in Codex, Claude Code, OpenCode, Hermes Agent and Pi that blocks four unrequested behaviors: scope creep, unneeded defensive code, exceeding a stated intent such as review-only, and repeating work that already has an answer.

• Agent skills

A symbolic framework for numeric optimization that implements forward- and reverse-mode automatic differentiation on sparse matrix-valued computational graphs, usable from C++, Python, MATLAB/Octave, Julia or JavaScript.

• Libraries

A command-line compiler that converts OpenAPI descriptions between JSON, YAML, and an equivalent Protocol Buffer representation, with a plugin interface for code generation and other API tooling.

• APIs

A unified, open benchmark and framework for building, orchestrating and evaluating domain-specific AI agents for Industry 4.0 asset operations and maintenance, with 140+ scenarios across 9 asset classes and specialist agents for IoT, failure modes and work orders.

• Agent frameworks
• AI evals

A hands-on guide, in four notebook sections, to building an AI agent from first principles up to an 'ambient' email assistant connected to the Gmail API, covering agent basics, evaluation, human-in-the-loop review and memory.

• Tutorials

Docker images and startup scripts for running the vLLM inference engine on NVIDIA DGX Spark hardware, from a single node to multi-node clusters over InfiniBand or RDMA.

• Inference
• Containers

A Rust-based, local-first web content extraction tool that turns a URL into clean Markdown, JSON or other LLM-ready output for AI agents and RAG pipelines, available as a CLI, an MCP server, a REST API and a self-hostable server.

• Browser automation

A personal AI agent runtime that treats chat as just one interface onto a system of record for conversations, runs, tasks, approvals, memory, skills, jobs, tools, traces and audit events, reachable identically from a CLI, a web app, mobile and messaging bridges.

• Agent frameworks

A cloud-native, distributed vector database for similarity search over embedding vectors in AI applications, with hybrid vector and scalar filtering search, replication and elastic scale-out, and client SDKs for Python, Go, Java and Rust.

• Databases

An unofficial Emacs plugin for GitHub Copilot that adds inline ghost-text completions, an interactive chat interface and Next Edit Suggestions, talking to the official Copilot language server directly over JSON-RPC rather than through a generic LSP client.

• Editors

A self-hosted control layer for running many AI coding agents across a fleet of a user's own machines from one console, scheduling and isolating each agent in its own sandboxed workspace and restarting ones that stall.

• Coding agents

Infrastructure deployed between an LLM API and a front-end application that gives a stateless model persistent memory and awareness across sessions and devices, through a unified instruction protocol, layered memory storage, a plugin engine and multi-agent collaboration.

• Agent memory

An open-source, highly customizable agent workspace, with a high-performance Rust runtime and a desktop app, that offers four agent-harness modes for different kinds of work and keeps a task moving across desktop, mobile and browser.

• Agent frameworks
• Desktop

An extension of the veRL reinforcement-learning library for training large language and vision-language model agents, using a step-independent multi-turn rollout design that supports long-horizon, many-step tasks and a new algorithm called Group-in-Group Policy Optimization.

• Models

A zero-dependency C++ command-line tool and MCP server that maps a codebase for an AI coding agent before it reads the repository, then checks a proposed edit's blast radius, the tests that reach it and where its quality got worse.

• Agent memory
• Code quality

A large-language-model multi-agent framework, originating from AntGroup's financial business practice, providing reusable multi-agent collaboration patterns such as plan-execute-express-review and letting a team fold its own domain expertise into how agents work together.

• Agent frameworks