• 2
• 0
• 0

HunyuanWorld-Voyager

Voyager is an interactive RGBD video generation model conditioned on camera input, and supports real-time 3D reconstruction.

• 2
• 0
• 0

multi-agent-coding-system

Reached #13 on Stanford's Terminal Bench leaderboard. Orchestrator, explorer & coder agents working together with intelligent context sharing.

• 2
• 0
• 0

osaurus

Native, Apple Silicon–only local LLM server. Similar to Ollama, but built on Apple's MLX for maximum performance on M‑series chips. SwiftUI app + S...

• 2
• 0
• 0

USO

🔥🔥 Open-sourced unified customization model

• 3
• 2
• 0

spec-kit

💫 Toolkit to help you get started with Spec-Driven Development

• 2
• 1
• 0

Youtu-agent

A simple yet powerful agent framework that delivers with open-source models

• 2
• 0
• 0

Flash-Sparse-Attention

🚀🚀 Efficient implementations of Native Sparse Attention

• 2
• 0
• 0

dionysus

laptop

• 2
• 1
• 0

vipe

ViPE: Video Pose Engine for Geometric 3D Perception

• 2
• 3
• 0

InfiniteTalk

​​Unlimited-length talking video generation​​ that supports image-to-video and video-to-video generation

• 2
• 0
• 0

mobile-use

AI agents can now use real Android and iOS apps, just like a human.

• 2
• 0
• 0

rendergit

Render any git repo into a single static HTML page for humans or LLMs

• 2
• 0
• 0

memU

MemU is an open-source memory framework for AI companions

• 2
• 1
• 0

deepagents_1

• 2
• 0
• 0

syncora-sdk

• 2
• 0
• 0

dinov3

Reference PyTorch implementation and models for DINOv3

• 2
• 0
• 0

ScreenCoder

ScreenCoder — Turn any UI screenshot into clean, editable HTML/CSS with full control. Fast, accurate, and easy to customize.

• 2
• 1
• 0

WeKnora

LLM-powered framework for deep document understanding, semantic retrieval, and context-aware answers using RAG paradigm.

• 2
• 3
• 0

deepagents

• 2
• 0
• 0

claude-code-security-review

An AI-powered security review GitHub Action using Claude to analyze code changes for security vulnerabilities.