Explore the directory

A curated catalog and dataset of 50 cloneable open-source GenAI applications and Hugging Face Spaces spanning RAG, AI agents, LangGraph, multi-agent systems and MCP, ranked for cloneability and setup clarity.

• Curated lists

A vLLM plugin that replaces a language model's dense classification head with a two-stage clustered retrieval pipeline, cutting scored vocabulary tokens roughly tenfold for faster, training-free inference.

• Inference

A Triton-based operator library providing optimized vLLM-specific kernels such as MoE routing and fused expert operators, used alongside the general FlagGems operator library for faster LLM inference.

• Inference

An MLX port of training-free inference speedups for Dream and DiffusionGemma diffusion language models on Apple Silicon, with dual-cache decoding and benchmark scripts.

• Inference

A self-hosted browser workspace that gives persistent, split-pane terminal access to Claude Code, Codex, OpenCode and a plain shell on a Linux server, keeping sessions alive in tmux across disconnects.

• Coding agents
• Terminals

A local web chat interface that connects LM Studio and other OpenAI-compatible model providers to an in-app tool runtime and the Supertonic text-to-speech engine, with hybrid memory search across conversations.

• Inference
• Speech

A mixture-of-experts model compression toolkit that prunes, merges and quantizes large MoE language models into smaller specialist models for local inference, implementing REAP expert pruning with several scoring strategies and a self-contained evaluation harness.

• Models
• Inference

A cross-platform Tauri desktop app that gives coding agents such as Claude Code, Codex, OpenCode and Cursor Agent live access to real SSH terminal sessions over MCP, while keeping passwords, private keys and sudo prompts entirely out of the AI's path.

• Coding agents
• Terminals

An Electron desktop workspace for creating, running and coordinating AI agents ('minds') powered by GitHub Copilot, built around Agent-to-Agent collaboration: minds discover each other, message each other, delegate stateful tasks and hand off artifacts.

• Agent frameworks
• Desktop

A token-native knowledge base for LLM agent memory that keeps storage, indexing, search and graph traversal entirely in discrete token-ID space, decoding text only at the final retrieval step, avoiding embedding APIs and vector databases altogether.

• Agent memory
• RAG

A memory layer for AI assistants that tracks what a user said, when, and whether it still lines up with what they are saying now, buildable as a bot's memory backend, a standalone chat app, a Discord bot, a REST API or an MCP server for Claude Code.

• Agent memory

An open-source visual canvas and runtime for building a governance and reliability control layer around an AI agent, compiling one flow specification to LangGraph, CrewAI, Mastra or the Microsoft Agent Framework and tracing every decision in Langfuse.

• Agent frameworks
• AI evals

A local-first, auditable memory control plane for AI agents that records exact host-observed text and verifiable execution evidence, authorizing and scoping what gets stored, injected, corrected or deleted, with integrations for OpenClaw, MCP, Pydantic AI and LangChain/LangGraph.

• Agent memory

A self-running pipeline of LLM agents that reads AI-related sources across Twitter, RSS, Hacker News, Reddit, Bluesky and arXiv, then publishes a nightly digest, a front page, a model-release timeline and a growing wiki through GitHub Actions.

• Agent frameworks
• Workflow automation

A single Rust binary that acts as an LLM-driven browser automation agent: it navigates, clicks and extracts data from a described goal in plain English, while also exposing the full DevTools surface as agent tools, plus a built-in MCP server.

• MCP servers
• Browser automation

Workshop and demonstration materials for an 'Architecting agentic AI solutions' course taught at NUS-ISS, including Jupyter notebooks for two graded workshops, their assignments, and a stripped-down starter version of one workshop for following along.

• Tutorials

A continuous-evaluation pipeline for multi-agent systems built with Google's Agent Development Kit and the A2A protocol on Cloud Run, deploying agents to shadow revisions and running automated evaluation suites to drive data-driven deployment decisions.

• AI evals

A library that stops one AI agent from silently overwriting another's concurrent edits to shared state such as a plan file or memory store, using a MESI-style coherence protocol and optimistic-concurrency commits verified in TLA+, across LangGraph, CrewAI and AutoGen.

• Agent frameworks
• Libraries

A replicable Google Cloud reference implementation of an enterprise-governed multi-agent system: a supervisor agent delegates to three specialist agents over the A2A protocol, with every model, MCP tool and agent call routed through Apigee for auth, quotas and token analytics.

• Agent frameworks
• APIs

Adds hotword-biased decoding to the WeNet C++ speech recognition runtime, using fuzzy phoneme matching and an automatic multi-objective tuner.

• Speech

A single-binary Git server with no database or leader, backed entirely by an S3 or GCS bucket, so any machine running it is a disposable cache of the repository.

• Git

Transfers files between two devices with no network, by animating one device's screen as a stream of QR codes for the other device's camera to read.

• Desktop utilities

A systems programming language in which the architecture — boundaries, protocols and proof obligations — is executable, and any implementation of it can be replaced.

• Languages

A self-hosted platform where an AI operator builds, deploys, runs, debugs and reruns data pipelines against a team's existing databases, warehouses and APIs.

• Data engineering
• Workflow automation