Explore the directory

A production inference stack that compiles a single model codebase to run on NVIDIA, AMD, Intel, TPU, or Trainium hardware without rewriting it for each accelerator.

• Inference

A lightweight reverse proxy and container orchestrator with a web UI for self-hosters, discovering Docker and Podman containers and routing traffic to them automatically.

• Hosting
• Networking

A Lua replacement for the copilot.vim plugin that brings GitHub Copilot suggestions and chat into Neovim with lower overhead and a native language server.

• Editors

A lightweight Python framework for building LLM-powered applications by setting up multiple agents, equipping them with models, vector stores and tools, and having them collaboratively solve tasks by exchanging messages.

• Agent frameworks

A Claude Code plugin that can run the same research, design or coding task through up to twelve external AI providers in parallel and flags disagreements before the work ships.

• Coding agents

A system for processing large collections of structured and unstructured data with LLMs, letting users describe operations in natural language while it orchestrates, optimizes and parallelizes the pipeline.

• Data engineering

A fast tokenizer for language models, written in Rust, that aims to match HuggingFace Tokenizers output while running at multi-gigabyte-per-second throughput.

• Libraries

An open-source platform for managing large language models and their related assets, such as datasets, spaces and code, with upload, download, storage, verification and distribution through a web interface, git, a chatbot or an SDK.

• Models

An open-source system of Chinese pretrained language and multimodal models, led by IDEA Research Institute, spanning general, generation, and domain-specific model families.

• Models

A fast Python library for serializing and validating data as JSON, MessagePack, YAML, and TOML using standard type annotations.

• Libraries

A desktop app that records an on-screen work session, uses the GitHub Copilot CLI to reconstruct it as an intent and ordered steps, and turns the result into a reusable skill or scheduled automation.

• Agent skills
• Workflow automation

A production-oriented Java platform for building agentic retrieval-augmented generation applications, covering document ingestion, hybrid retrieval, intent understanding, session memory and MCP tool calling.

• Agent frameworks
• RAG

A CLI and MCP server that points a headless browser at a live website and extracts its design system as design tokens and framework-ready files.

• MCP servers
• Design tools

An open-source collaborative workspace where multiple AI agents connect, share context and work together from a single browser-based surface, with a CLI and desktop app for managing them.

• Agent frameworks

A free course teaching how to apply Transformers to natural language processing and other tasks using the Hugging Face ecosystem, including its Transformers, Datasets, Tokenizers and Accelerate libraries.

• Tutorials

An open-source document intelligence API that performs layout analysis, OCR and semantic chunking to convert PDFs, presentations, Word documents and images into structured data for RAG and LLM pipelines.

• RAG

A pre-alpha multi-agent orchestration platform that generates task DAGs and coordinates specialized agents for research, coding, design and unattended workflow automation.

• Agent frameworks

An open-source natural language processing library built on Apache Spark, offering pretrained pipelines and models for tasks like tokenization, named entity recognition, sentiment analysis, translation and question answering.

• Libraries
• Data engineering

A Kotlin and Java API for generating .kt Kotlin source files programmatically.

• Libraries

An open-source multi-agent collaboration platform where each AI agent has a persistent identity, long-term memory and its own workspace, functioning together as a crew of digital employees within an organization.

• Agent frameworks

A command-line tool that bulk port-forwards every Kubernetes service in a namespace to a unique local loopback IP, so local code reaches them by their in-cluster service names.

• Containers
• Networking

An open-source, modern AI training tracking and visualization platform that integrates with more than 50 frameworks such as PyTorch, Transformers and Ultralytics, usable as a cloud service or self-hosted, including on Kubernetes.

• Models

A self-hostable, single-binary workflow engine for operations and internal automation, defining DAGs in declarative YAML with scheduling, retries, human tasks, and a built-in web UI.

• Backend
• Workflow automation

An open-source, self-hosted development workspace that persists and improves across sessions, running unattended multi-step tasks and scheduled jobs from a desktop app, web dashboard, CLI, or chat tools such as Slack and Discord.

• Coding agents