vLLM: High-Throughput LLM Inference Engine with PagedAttention
Introduction vLLM is a high-throughput and memory-efficient inference and serving engine for large language models that has fundamentally transformed how organizations deploy LLM inference at scale. Originally developed at UC Berkeley’s Sky Computing Lab, vLLM was published at ACM SOSP...
SUMO-RL: Reinforcement Learning for Traffic Signal Control with OpenAI Gym
SUMO-RL: Reinforcement Learning for Traffic Signal Control with OpenAI Gym SUMO-RL reinforcement learning traffic simulation brings the power of modern RL algorithms to urban traffic signal control by wrapping the SUMO (Simulation of Urban MObility) simulator into standard Gymnasium and...
Local Deep Research: Open-Source AI Research Assistant Achieving 95% SimpleQA...
Introduction Local deep research has emerged as one of the most significant breakthroughs in open-source AI tooling, and the Local Deep Research project by LearningCircuit is leading the charge. With 7,243+ GitHub stars and growing at over 2,400 stars per...
Eclipse SUMO: Microscopic Multi-Modal Traffic Simulation Platform
Introduction Eclipse SUMO (Simulation of Urban MObility) is a leading open-source platform for microscopic traffic simulation, enabling researchers, engineers, and urban planners to model complex transportation networks with unprecedented fidelity. Developed by the German Aerospace Center (DLR), Institute of Transportation...
Pi Mono: The Full-Stack AI Agent Toolkit From libGDX Creator Mario Zechner
Pi Mono: The Full-Stack AI Agent Toolkit From libGDX Creator Mario Zechner The pi-mono AI agent toolkit has rapidly become one of the most starred open-source projects in the AI coding space, accumulating over 39,132 stars on GitHub. Created by...
CyberVerse: Open-Source Digital Human Agent Platform with Real-Time Video Cal...
CyberVerse: Open-Source Digital Human Agent Platform with Real-Time Video Calling CyberVerse is an open-source digital human agent platform that transforms a single photograph into a living, breathing AI character you can talk to face-to-face in real time. Unlike pre-recorded avatars...
UI-TARS Desktop: ByteDance's Open-Source Multimodal AI Agent That Controls Yo...
UI-TARS Desktop: ByteDance’s Open-Source Multimodal AI Agent That Controls Your Computer ByteDance’s UI-TARS-desktop is a multimodal AI agent stack that lets you control computers and browsers through natural language, powered by Vision-Language Models (VLMs). With over 31,000 stars on GitHub...
Symphony: OpenAI's Coding Agent Orchestrator for Autonomous Work
Symphony: OpenAI’s Coding Agent Orchestrator for Autonomous Work OpenAI’s Symphony is a specification-first, language-agnostic orchestration service that turns project work into isolated, autonomous implementation runs for coding agents. Rather than having engineers supervise individual coding agent sessions, Symphony manages the...
Solar Map: Interactive Sunshine Intensity Map and Solar Panel Optimizer
Solar Map: Interactive Sunshine Intensity Map and Solar Panel Optimizer The Solar Map is an interactive web application available on PyShine that combines real-time sun position tracking with solar panel energy estimation. Built with Leaflet.js and SunCalc, it provides a...
RAGFlow: Open-Source RAG Engine with Deep Document Understanding
RAGFlow: Open-Source RAG Engine with Deep Document Understanding RAGFlow is a leading open-source Retrieval-Augmented Generation engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs. With nearly 80,000 GitHub stars and a thriving community,...
PageIndex: Vectorless Reasoning-Based RAG That Achieves 98.7% on FinanceBench
PageIndex: Vectorless Reasoning-Based RAG That Achieves 98.7% on FinanceBench PageIndex is a vectorless, reasoning-based RAG system that replaces vector similarity search with LLM reasoning over hierarchical tree indexes, achieving state-of-the-art 98.7% accuracy on FinanceBench. Traditional RAG pipelines rely on embedding...
Gemma Chat: Local AI Coding Agent for Apple Silicon via MLX
Gemma Chat: Local AI Coding Agent for Apple Silicon via MLX Gemma Chat is a local AI coding agent that runs entirely on Apple Silicon via Apple’s MLX framework, enabling fully offline vibe coding without API keys, cloud services, or...