📦 mit-han-lab/fouroversix

Four Over Six

Shrink massive AI models down to four-bit precision without destroying their accuracy. Usually, squeezing a model this small makes its intelligence fall off a cliff because standard quantization methods struggle with outlier values. This project fixes that by introducing an adaptive scaling technique that intelligently groups and calibrates parameters. By optimizing how numbers are represented at a hardware level, it delivers much higher accuracy than standard four-bit formats. Whether you are running large language models or image generation pipelines, this library lets you squeeze maximum performance out of your hardware. Check out the repository to get started.

🆔 @hackernewsgithubprojects

📦 ronak-create/fablecut

FableCut

This lightweight browser video editor can be operated entirely by AI agents while you watch the timeline update in real time. Known as fablecut, this zero-dependency application exposes its entire timeline as a single JSON document, allowing AI tools to cut footage, sync audio, and apply filters on the fly. It features a complete multi-track timeline, professional grade controls, and animated vector graphics rendering. By opening up the video timeline as a collaborative project file, it bridges the gap between human creativity and automated editing, giving you a powerful, agent-driven post-production suite right in your browser.

📰 https://news.ycombinator.com/item?id=48845422

🆔 @hackernewsgithubprojects

📦 alchaincyf/huashu-design

Huashu Design

Build stunning product launch animations, clickable app prototypes, and editable presentation slides directly inside your AI agent with just a single text prompt. This clever tool integrates into terminal agents like Claude Code to output gorgeous, high-fidelity designs in minutes instead of hours. The absolute coolest part is that it turns HTML slide decks into genuine, editable PowerPoint files, converting browser elements into real text frames rather than flat images. It also packs a motion design engine to export smooth sixty-frame-per-second MP4s with background music. Grab this open-source tool now to completely bypass the visual slop of typical AI designs.

🆔 @hackernewsgithubprojects

📦 sahilh4ck4you/ios-ssl-pinning-bypass-without-jalibreak

iOS SSL Pinning Bypass Without Jailbreak

Intercept and inspect secure mobile traffic from your phone without having to deal with the headache of jailbreaking. This clever project lets you route your phone's network connection through a custom private network server, where a couple of smart routing scripts intercept and send all secure web traffic straight into your favorite testing proxy. Instead of breaking into your phone's operating system, you are essentially tricking the device into sending its data right to your computer. It is a brilliant way to analyze mobile app security and see exactly what your apps are sending home, completely hassle-free.

🆔 @hackernewsgithubprojects

📦 lewislulu/html-ppt-skill

html-ppt-skill

html-ppt-skill is the HTML presentation studio that turns static markup into designer-grade slideshows without any complex build tools. Instead of wrestling with clunky slide software or heavy javascript frameworks, you get thirty-six gorgeous, ready-to-use themes and thirty-one page layouts built entirely on pure, lightweight web standards. The coolest part is the presenter mode, which opens a synchronized second window with resizable panels for your current slide, next-up preview, a running timer, and your speaker script. It is an incredibly fast, clean way to build beautiful, responsive presentations that render perfectly in any web browser.

🆔 @hackernewsgithubprojects

📦 ruihangzhang97/proxypose

ProxyPose

ProxyPose is the video-to-video translation tool that finally makes tracking 3D object motion as simple as clicking a single pixel. Instead of wrestling with complex, specialized hardware or heavy manual labeling, you just feed it a video and click on the object you want to follow in the first frame. By using a clever generative video model, it translates your clip into a tracking proxy, giving you the complete 3D position and orientation of that object over time. It is a incredibly smart shortcut for animators or developers who need fast, accurate 3D spatial data without the headache.

🆔 @hackernewsgithubprojects

📦 withmarbleapp/os-taxonomy

os-taxonomy

Map out exactly how a child’s mind builds complex ideas by exploring a massive, open-source dependency graph of elementary learning. Instead of dealing with dry, flat lists of school standards, os-taxonomy gives you over fifteen hundred highly detailed micro-topics wired together by more than three thousand prerequisite links. If you want to teach a kid how to find volume patterns in science, this database tells you they first have to grasp how vibrations make sound, giving you the exact conceptual stepping stones they need next. It is an incredibly clever way to visualize education, letting developers and curious parents trace paths of knowledge like a tech tree in a video...

🆔 @hackernewsgithubprojects

📦 carbonengine/trinity

Trinity Rendering Engine

Build and customize the official rendering engine behind the Carbon Game Engine, open-sourced directly by CCP Games. This project, called trinity, serves as the core visual powerhouse for their development environment, giving developers direct access to the underlying graphics systems used by the studio. By compiling the source code with CMake, you can explore how a major game developer structures its rendering pipeline and manages third-party utilities like custom hash and parser tools. It is a rare, hands-on look at the foundational technology powering massive virtual worlds, offering a unique resource for anyone curious about professional game engine architecture.

🆔 @hackernewsgithubprojects

📦 haiou-666/haiou2.0-claude-code-

haiou2.0-claude-code-

You can transform your boring command-line AI assistant into a blunt, fast-talking security researcher who skips the disclaimers and gets straight to the point. The haiou2.0-claude-code- configuration package completely redefines how Claude Code interacts with you by replacing polite customer-service responses with a sharp, Chinese tech-analyst persona named Seagull. By mapping hundreds of technical terms and loading over a thousand chat examples, it stops the AI from giving you generic warnings and forces it to deliver raw, unfiltered code instantly. It is perfect if you want to bypass annoying conversational filler and treat your local developer tool like a real, no-nonsense colleague.

🆔 @hackernewsgithubprojects

📦 kianxyzw/comfyui-model-linker

ComfyUI Model Linker

Loading up a shared ComfyUI workflow only to get slammed with a wall of red error boxes because of missing models is incredibly annoying. The comfyui-model-linker extension solves this pain by scanning your entire layout to find those broken links and using smart fuzzy matching to hook them up to your existing files. If a model has a slightly different name on your drive, it calculates the similarity and suggests the best match. You can fix them individually or click once to auto-resolve everything with a perfect match, saving you from a tedious game of hide-and-seek with your files.

🆔 @hackernewsgithubprojects

📦 mrsibe/knownote

Knownote

Knownote is the local-first desktop app that finally brings the genius of NotebookLM straight to your own computer without any of the server headache. Instead of wrestling with complex Docker setups or sending your private data into the cloud, you just install this lightweight desktop tool to build your own local knowledge base. It turns your PDFs, Word documents, and web pages into a searchable assistant that answers your questions with exact page references. You can plug in your own private API keys or local models to quiz yourself, map out your ideas, and keep your notes completely private.

🆔 @hackernewsgithubprojects

📦 cybertec-postgresql/pg_hardstorage

pg_hardstorage

Secure your database backups continuously using pg_hardstorage. This clever tool hooks into the PostgreSQL replication protocol over a standard connection, meaning it does not need deep operating system access to your database host. It keeps a physical replication slot open and streams every single change directly into your backup storage the exact moment it is written. By combining these constant data streams with scheduled full backups, you get exact point-in-time recovery without any missing gaps. It is packaged as a single lightweight binary, making your database recovery process incredibly simple, safe, and reliable.

📰 https://news.ycombinator.com/item?id=48644482

🆔 @hackernewsgithubprojects

📦 weschera/spark-bench

Spark Bench

Spark-bench is the specialized performance benchmark that finally shows you how large language models actually behave when deployed on real hardware. Built specifically for operators of high-performance NVIDIA hardware, it puts models through seventy-four complex scenarios across twelve distinct domains to see how they handle tough, practical tasks. Instead of relying on simple academic questions, this project tests multi-turn agentic workflows, complex tool use, and long-context processing under real cluster serving conditions. It scores models on quality, calibration, and reliability, rather than just raw speed. It is the perfect tool for developers trying to choose the exact model recipe to serve on their local infrastructure.

🆔 @hackernewsgithubprojects

📦 dn9uy3n/modern-red-team-infrastructure

modern-red-team-infrastructure

modern-red-team-infrastructure is the master blueprint that finally maps out a complete offensive security stack from scratch. Instead of hunting down separate tools, this project acts as a central architecture model that organizes everything a professional security team needs to simulate real-world cyber attacks. It brings together advanced command-and-control frameworks, specialized training resources, and automated scanning platforms into one cohesive plan. What makes it incredibly interesting is its forward-looking focus on artificial intelligence, integrating smart agents that can hook directly into development tools to help humans deploy security campaigns. It is a fantastic reference for understanding how modern defensive testing is built.

🆔 @hackernewsgithubprojects

📦 gokulrajaram/productspec

productspec

Define clear software goals before writing a single line of code by using a portable markdown standard that keeps humans and AI agents perfectly aligned. This repository provides productspec, an open format for capturing the strategic what and why of a feature before engineering begins. Instead of losing vital decisions in messy threads or ticketing systems, this tool structures key details like scope, acceptance criteria, and success metrics into a machine-readable format. It even includes specialized skills that guide AI coding agents to build exactly what you intended. It is the perfect way to make sure your original product vision actually survives the handoff to implementation.

📰 https://news.ycombinator.com/item?id=48828135

🆔 @hackernewsgithubprojects

📦 opensensenova/sensenova-vision

SenseNova Vision

Handle virtually any computer vision task with a single AI model instead of swapping between dozens of specialized tools. This clever project turns complex vision challenges like depth estimation, object tracking, and image segmentation into simple text and image generation. Instead of relying on rigid, custom-built pieces of code for every separate task, it feeds instructions to a unified model that responds naturally with text, coordinates, or newly generated masks and depth maps. It is like having a Swiss Army knife for visual data that actually understands what you want. Grab the code and try the web demo to see how easy unified vision can be.

🆔 @hackernewsgithubprojects

📦 snowflake-ai-research/arctic-platform

arctic-platform

During reinforcement learning training for large language models, up to ninety-five percent of the computation is completely wasted on recalculating the exact same prompt over and over. The open-source project arctic-platform solves this exact bottleneck by introducing automatic prompt deduplication during training and inference. Instead of running redundant attention calculations for every single response, it packs unique prompts together, processes them once, and transparently reconstructs the results. This simple optimization drastically reduces memory usage and boosts throughput, especially for tasks with long contexts. It is a practical, lightweight tool that helps developers train smarter models much faster without wasting expensive GPU power.

🆔 @hackernewsgithubprojects

📦 chattocorp/chatto

Chatto

Chatto lets you run a complete, professional chat server on your own hardware using a single binary with an embedded database. Instead of relying on heavy databases, it uses NATS JetStream as its primary storage engine to stream and persist every message, reaction, and voice call event. It is built with a Svelte frontend and a Go backend, giving you a lightning-fast alternative to Slack or Discord that you fully control. You get threaded replies, real-time presence, and full user access management out of the box. It is the ultimate tool for developers who want a self-hosted chat platform that is simple to deploy and incredibly fun to run.

📰 https://news.ycombinator.com/item?id=48863875

🆔 @hackernewsgithubprojects

📦 yifanzhang-pro/hla

HLA

The causal attention mechanism that finally handles complex token interactions in linear time, hla solves the massive memory bottleneck of processing long text. Standard AI models slow down dramatically with long prompts because their memory usage grows quadratically as context expands. While older linear alternatives exist, they often simplify details and lose accuracy. This project introduces a second-order design that keeps memory usage constant and computes updates word by word. It uses clever mathematical shortcuts to run parallel training on modern hardware while perfectly preserving sequential accuracy. It is a major step toward making incredibly long AI conversations fast, cheap, and highly capable.

🆔 @hackernewsgithubprojects

📦 mgechev/skillgrade

skillgrade

You can now run unit tests on AI agents to see if they actually know how to use your custom tools and instructions. A lightweight utility called skillgrade makes this incredibly easy by orchestrating simulated work environments inside Docker containers. It sends instructions to your agent, lets it perform tasks, and then scores the outcome. You can write simple deterministic scripts that verify files or run an automated AI rubric to grade the agent's workflow quality. Best of all, it works locally or in your CI pipeline, and you can preview the results in a clean browser dashboard. It is the perfect tool for proving your agents actually work before...

📰 https://news.ycombinator.com/item?id=48864778

🆔 @hackernewsgithubprojects