📦 nvlabs/motionbricks

MotionBricks: Real-Time Character Animation

Generate fluid, real-time motion for human characters and robots by simply placing a handful of keyframes. MotionBricks is a neural model that intelligently fills in the missing movement between those poses, creating smooth transitions on the fly. Its standout feature is the use of Smart Primitives, which turn specific actions into reusable building blocks. This allows developers to assemble complex behaviors like vaulting or sitting without retraining the entire system. By chaining these generations together, the software drives continuous locomotion and object interactions directly on an NVIDIA GPU. It bridges the gap between keyframe animation and dynamic simulation.

🆔 @hackernewsgithubprojects

📦 ahtx/kindi

kindi

A single gigabyte graphics card from ten years ago just ran a modern twenty-seven billion parameter language model. Kindi is a custom inference engine built specifically for low cost NVIDIA hardware like the Quadro K4220 and K1200. It runs these older cards at speeds that are thirteen times faster than standard tools, making old hardware actually useful for running local AI. Instead of needing a new expensive rig, you can plug in cheap used cards and get real performance. This project solves the cost problem by making powerful AI accessible on a budget.

🆔 @hackernewsgithubprojects

📦 tuxkoh/ninfer-v100xn

NInfer: Run Large Language Models on an Old GPU

Run massive language models on a single, aging graphics card. NInfer is a specialized inference engine designed to squeeze serious performance out of older Nvidia hardware, specifically the V100. It skips the usual bulky frameworks and instead writes custom code directly for the GPU's specific architecture. This means you can deploy large AI models without needing a multi-thousand dollar server or the latest, expensive chips. It is a fascinating look at how much optimization can replace raw hardware power.

🆔 @hackernewsgithubprojects

📦 tencent-hunyuan/hy-simulmt

Hy-SimulMT Translates As You Talk

Hy-SimulMT is the real-time translator that actually keeps up with you. It takes audio as you speak and translates it into speech instantly, so you never have to wait for a pause. The coolest part is that it clones the speaker's voice, meaning you hear the translation in the original person's tone. It works back and forth between Chinese and English. This removes the awkward silence from cross-language chats. It’s perfect for meetings where speed matters. No more waiting for the interpreter to catch up. It just follows along, keeping the conversation flowing naturally. If you talk across languages, this changes the game.

🆔 @hackernewsgithubprojects

📦 cryogenicplanet/arc

Arc Mod: Use Arc Without an Account

Browses the web without asking for your Arc account or personal data. This project takes the popular Arc browser for Mac, strips away the mandatory sign-in, and blocks all traffic sent back to the company’s servers. You get the full browser experience with your privacy intact, plus a built-in feature that lets AI coding agents control their own hidden, isolated browser windows in the background. It runs on your machine as a separate app, meaning your original Arc install stays completely untouched. If you love the interface but refuse to hand over your personal info to a cloud service, this is your ticket to browsing privately.

🆔 @hackernewsgithubprojects

📦 whokilleddb/tts

TTs: The Hidden Shellcode Loader

TTs is the payload builder that hides executable code inside standard Visual Studio template files. This tool takes raw machine instructions, encrypts them, and wraps them into a legitimate text transform file. It then compiles a custom piece of code that loads this file. The most surprising part is that it uses the language runtime itself to run the code without any obvious memory allocation calls. The result is a small package that looks like a normal developer utility. This is a clever way to test how standard tools can be repurposed. It shows how much trust we place in familiar file types.

🆔 @hackernewsgithubprojects

📦 0c0de/splatterhouse-decomp

Splatterhouse Recompiled: Xbox 360 on PC

Splatterhouse Recompiled is the Xbox 360 port that finally runs a console game on your PC without an emulator. It takes the original game code from your disc and translates it directly into native instructions, meaning your computer executes the game at full native speed with no runtime interpretation slowing things down. This is why it hits a solid sixty frames per second and supports ultrawide monitors automatically. You don't need to install complex software; just drag your own game disc file onto the installer, and it handles the rest. This project proves that with static recompilation, older console titles can get a second life on modern hardware.

🆔 @hackernewsgithubprojects

📦 ranahaani/i-hate-editing

I Hate Editing

I Hate Editing is the video editing skill that refuses to hand back work that still needs fixing by hand. Instead of scrubbing a timeline, it listens to your raw audio, finds the silences, and cuts the video precisely where your voice pauses. The most surprising part is how it verifies those cuts by re-transcribing the audio right after the join to catch any accidental word repeats before you ever see a preview. It also captures real web pages as tall images, scrolling and zooming them to match your narration rather than using fake screenshots.

📰 https://news.ycombinator.com/item?id=49597350

🆔 @hackernewsgithubprojects

📦 lawrencewzen/wx-bot

Give your AI a WeChat Brain

wx-bot pipes your WeChat messages directly into AI agents like Grok, letting your personal assistant chat with you through the app you already use. It handles the messy backend plumbing, storing every message in a local database so nothing gets lost, even if your AI is busy. You can push updates instantly or let the AI pull them at its own pace, making it flexible for different setups. It’s a clean, secure bridge that keeps your data local and only sends traffic out, giving your AI a reliable, private line to your real-life conversations. It’s a clever, practical way to make your digital helper feel actually present.

🆔 @hackernewsgithubprojects

📦 diaoorg/awesome-supermimic

How Awesome SuperMimic Makes Robots Dance

Awesome SuperMimic is the community hub that tracks how researchers teach robots to dance. Built around a Berkeley paper that uses a learned world model to improve movement policies, it curates thirty-three projects where humans actually put these algorithms on real hardware. The most fascinating part is seeing independent labs reproduce the results, with some teams getting robots to perform flawless backflips in under two hours. If you are tracking the future of humanoid control, this list is the fastest way to find what is working and where the technology is heading.

🆔 @hackernewsgithubprojects

📦 vibha-ramprakash/claude-video-effects

Claude Video Effects: 45 Talking-Head Edits in One Command

Drop a talking-head clip into Claude Code and type the name of an effect to render a finished video in seconds. That is the whole workflow for a toolkit built by Vibha Ramprakash that turns plain phone footage into polished social media content. The real magic is face tracking. Instead of manually keying every frame, the system locks onto your face and places animations like orbiting reels or glowing widgets directly behind your head, keeping you perfectly visible while the background does the work. It handles the complex tracking and rendering for you, so you just say what you want and watch the result appear.

🆔 @hackernewsgithubprojects

📦 musictribeopensource/0700-aam

The AI-Driven Hardware DAW: 0700-AAM

Turn a standard beat machine into a gesture-controlled studio by flashing the Behringer KICK with its open source firmware. This project provides the actual code for the device, letting you see exactly how its AI camera tracks your hands to trigger sounds. It also reveals the audio engine powering its sampler and effects. You are no longer limited by closed hardware; you can modify, debug, or extend the music creation tools in your hands. This transparency builds a community around a tangible instrument. It is a rare look at the digital soul of a physical synthesizer. Now you have the blueprint. Go build something new with it.

🆔 @hackernewsgithubprojects

📦 slopus/happy-desktop

Happy Desktop: One App for All AI Agents

Orchestrate Claude, Codex, and Grok inside a single desktop window. Happy Desktop solves the headache of juggling different apps for each coding agent by letting you run them all side by side. The real surprise here is that it is natively multiplayer. You can invite a friend into your active session so you both watch the agent work in real time and steer it together. It is free, open source, and runs on your own hardware, meaning your code stays exactly where you put it. No extra bills, just a smarter way to manage your existing subscriptions.

🆔 @hackernewsgithubprojects

📦 joshuaswarren/metal2vk

metal2vk: Run Apple's GPU Code on Linux

metal2vk is the translator that finally lets you run Apple’s GPU code on Linux. If you’ve ever built high-speed AI software using Apple Silicon, you know the catch: the speed comes from custom kernels written in a language that only works on Mac. This project fixes that dead end by reading those native files and converting them into a universal format that any modern GPU can understand. No rewriting, no new coding, just a clean bridge. It uses open-source tools to handle the heavy lifting, meaning you get the raw performance of the original hardware without needing Apple’s own software.

🆔 @hackernewsgithubprojects

📦 hkuds/autoswarm

AutoSwarm: The Virtual Agency

A single command can spin up a full digital company where distinct AI agents act like real employees. AutoSwarm solves the isolation problem by connecting different coding tools into one unified team. It assigns specific roles, from backend architecture to design, allowing them to collaborate on complex projects simultaneously. Instead of juggling separate chat windows, you manage a single workflow where these agents share context and divide the labor. This creates a structured environment for building software faster. The project turns individual assistants into a coordinated workforce, giving you a practical way to scale your development output without hiring a human team.

🆔 @hackernewsgithubprojects

📦 nvidia/auto-ontology

Auto Ontology: The Missing Brain for Your Data

Stop guessing how your data actually connects. NVIDIA released Auto Ontology, a framework that builds a living map of your database so AI agents never hallucinate a metric again. It takes your messy, scattered data definitions and turns them into a single source of truth, meaning when an agent asks for revenue, it gets the exact, verified number every single time. No more vague answers or silent errors, just clear, trustworthy insights you can actually trust. If you are tired of AI guessing at your numbers, this is the bridge you need.

🆔 @hackernewsgithubprojects

📦 synacktiv/livepyre

Livepyre: The Livewire RCE Tool

Execute arbitrary code on vulnerable Livewire applications with Livepyre, a specialized exploit tool built by Synacktiv. This repository addresses CVE 2025 54068, allowing attackers to run system commands if they know the application key or can bypass it through snapshot analysis. The single most striking capability here is its ability to achieve remote code execution even without the secret key, by intelligently scanning for object types to avoid brute force attempts. It turns a specific security flaw into a direct command line action, proving how dangerous exposed metadata can be. For security professionals, it highlights the critical need to protect application keys and patch dependencies immediately.

🆔 @hackernewsgithubprojects

📦 jiwoochris/artex-global

Artex Global: The AI Pentester

Artex Global lets you deploy a team of artificial intelligence agents to run autonomous penetration tests on your software. While you watch the dashboard, these agents independently break down targets, execute real security tools, and map out vulnerabilities in a live graph. The coolest part? It doesn't just find holes; it generates the exact detection rules you need to stop these specific attacks from happening again. This makes it a powerful asset for security teams who want to understand how modern AI-driven intrusions actually work. If you want to see how autonomous cyber defense plays out in real time, Artex Global is the project to watch.

🆔 @hackernewsgithubprojects

📦 twigpine/zero

Zero: The AI Coding Agent You Own

Take full control of your coding workflow with Zero, a terminal-based AI agent that runs entirely on your own machine. Unlike cloud-bound tools that lock you into specific providers, Zero connects directly to any model you choose, whether that is OpenAI, Anthropic, or a local option running on your hardware. It inspects your code, edits files, and executes commands while respecting strict permission boundaries you define. This means your data stays local, your sessions remain private, and you decide exactly what the agent is allowed to touch. It is a practical tool for developers who want powerful automation without surrendering their infrastructure or their data sovereignty.

🆔 @hackernewsgithubprojects

📦 samnet-dev/paqctl

paqctl: Bypassing the Great Firewall with Broken Packets

Send your internet traffic through a wall that is actively trying to block you. Paqctl is a tool that lets you connect to a free server from behind heavy censorship, like China's Great Firewall. The cool part is how it works. It sends data packets that are intentionally broken on purpose. Because these packets look wrong to the firewall, the system lets them pass right through. Inside that broken shell, your real data travels safely to the server. This means you can browse the open web even when everything else is blocked. It is a clever trick that turns the firewall's own rules against it.

🆔 @hackernewsgithubprojects