📦 jev-chat/jev-chat-jarvis-mac

Jev Chat Jarvis for Mac

A tiny app on your Mac reads the chat window in front of you without touching the database or taking screenshots. It uses a local model to guess what a message really means and how risky it is, then drafts a few replies you can paste into the chat. This is Jev Chat Jarvis for Mac, a floating helper that stays out of your way while you decide what to send. It is strictly read-only, so your privacy stays yours. The real surprise is that it runs entirely on your computer, so your conversations never leave your machine unless you choose to send them.

🆔 @hackernewsgithubprojects

📦 joeseesun/qiaomu-download

The Video Downloader That Actually Checks Your Files

Qiaomu Download is the video downloader that refuses to call a job done until the file actually plays. Most tools just grab a file and hope for the best, but this one runs a deep check to make sure your video isn't corrupted before you even know it arrived. It handles links from YouTube, TikTok, and even Spotify tracks by finding the matching official audio, all while keeping your private login data locked down. If you are tired of guessing whether a download is good, this gives you a clear green light so you know your media is ready to use.

🆔 @hackernewsgithubprojects

📦 kryvora-network/kryvora-node

Kryvora Node

Your computer is now a paid participant in a decentralized physical network. Kryvora Node runs a background daemon that connects your hardware to a central coordination hub, allowing it to run verification tasks and prove its availability to earn rewards. It is the official client for a DePIN network, handling everything from telemetry probes to peer state checks. To get started, you need a specific on-chain credential called a Genesis Node Key to authenticate your worker. Once connected, your machine becomes part of a distributed system that verifies real-world infrastructure, turning idle compute resources into active, verified nodes.

🆔 @hackernewsgithubprojects

📦 kydlikebtc/awesome-jev

What Is Awesome Jev?

Discover how real developers are using TypeSafe AI's Jev to solve specific decision problems. Instead of generic articles, this collection maps one thousand two hundred and seven public examples to exact decision patterns like intent routing or safety gating. You can see precisely how each project handles the logic, complete with source citations and link checks to verify accuracy. It also includes a compatibility guide showing you exactly how to switch between different providers without breaking your code. If you are building agents that need to make reliable, structured decisions, this is the field guide you need to see it done right in practice.

🆔 @hackernewsgithubprojects

📦 lsposed/zygisknext

Zygisk Next

Zygisk Next is the Android customization utility that finally frees you from relying on Magisk's built-in tools. If you love modding your phone, you know how tricky it is to inject custom modules into core system processes. This tool acts as a standalone engine that brings powerful app injection capabilities to other root environments like KernelSU. It even supports specialized setups like Xiaomi's modern HyperOS Rust runtime, letting developers hook into apps without dealing with heavy Java layers. It is a sleek, hyper-focused tool for anyone who wants absolute control over how their custom Android modules run behind the scenes.

🆔 @hackernewsgithubprojects

📦 mabdulre9/electrical-engineering

Learn Electrical Engineering for Free

Grab a full university curriculum in electrical engineering without spending a single dollar. This repository maps out a complete self-taught path, pulling together free lectures from heavyweights like MIT, Stanford, and Yale. It handles everything from the foundational math you need to survive calculus, straight through to advanced topics like digital communications and power systems. The coolest part is that it isn’t just a random pile of links. It’s carefully ordered so you know exactly which video to watch next, complete with recommended textbooks to deepen your understanding. It removes the expensive tuition barrier, letting you learn the real theory behind how circuits, signals, and machines actually work.

🆔 @hackernewsgithubprojects

📦 mak5er/aircard-ios

AirCard-iOS: Customize Your iPhone Without Jailbreaking

AirCard-iOS is the iOS tool that lets you customize Apple Wallet cards and lock screens without jailbreaking. Most customization apps require a computer or a cracked phone, but this project runs entirely on your device using a clever local tunnel. The standout feature is its ability to swap the artwork on your digital transit cards instantly, making your Suica or other cards look brand new the moment you open your wallet. It also themes your passcode dialer with live previews, so you can pick exactly how those buttons look. Since it handles everything locally via pairing, you never need to tether your phone to a Mac.

🆔 @hackernewsgithubprojects

📦 minorun365/minorun-marp-skill

Minorun Marp Skill: The AI Slide Master

Minorun Marp Skill is the presentation toolkit that actually makes AI generated slides look professional. It stops the machine from cramming cluttered text onto every screen by teaching it how to design real figures and tell a coherent story. Instead of giving you a generic template, it provides a dark theme and specific rules for balancing illustrations so your deck looks like a human made it. If you are building slides with AI, this repository is the missing piece that turns a rough draft into a polished talk. You stop fighting the output and start trusting the design, giving you back your time to focus on the actual presentation.

🆔 @hackernewsgithubprojects

📦 newliver666/apk-reverse

apk-reverse: The Agent's Playbook for Android Reverse Engineering

apk-reverse hands you a strict operational playbook to reverse engineer Android apps without wasting hours on dead ends. It acts as a rigid procedure for AI agents, forcing them to stop and verify their environment before touching a single byte of code. The real magic is catching the sneaky repack failure where an app looks perfect but silently rejects every request because it derives keys from its signing certificate. This prevents you from shipping a broken build that lies to you. It turns guesswork into a verifiable process. If you modify apps, you need this guardrail.

🆔 @hackernewsgithubprojects

📦 johnheibel/claudeanimationbase

Claude Animation Base

Claude Animation Base is the starter kit that turns a chatbot into a hand-painted cartoon animator. Instead of just typing code, you talk to the AI about the story, and it builds a complete video for you. The star is a little character named Clawd, who has thirty-one acted emotions and a guide that teaches the model how to time scenes so they feel alive. It is the only tool where you can ask for a butterfly chase and get a finished, boiling-ink animation without writing a single line of code yourself.

🆔 @hackernewsgithubprojects

📦 pallavi-shekhar/ai-engineering-interview-questions-company-wise

Real AI Interview Questions From 35 Companies

A single document collects actual interview questions asked at thirty-five major tech firms, organized specifically by employer rather than by topic. This project solves the problem of guessing what a specific company actually tests, mapping out their unique focus areas like large language model internals or agent design. It turns scattered online reports into a clear roadmap for candidates targeting roles from AI engineer to machine learning engineer. By showing exactly which questions recur at places like OpenAI or Google, it removes the uncertainty from preparation. Treat this not as a script, but as a map of what each organization values, giving you a direct line into their hiring mindset.

🆔 @hackernewsgithubprojects

📦 miuuyy/astra-ares

Astra Ares: Smarter AI Coding

Astra Ares is the coding assistant that lets your AI think only as hard as it needs to. Instead of burning through massive token budgets on every single line of code, it uses a smart sidecar to decide exactly how much mental effort each step requires. It checks your task, asks a helper model how deep to think, and applies that effort on the fly. This means simple tasks breeze through quickly while complex problems get the full brainpower they deserve. It’s a clever way to cut costs without sacrificing quality.

🆔 @hackernewsgithubprojects

📦 rgem227/knoweldge-base

Knowledge Base: The MCP Powerhouse

The knowledge base that finally lets your AI agents read your actual files is called Knowledge Base. It is a self-hosted hub where you upload documents, organize them into folders, and manage team access, all from a single web interface. The standout feature is its built-in support for Model Context Protocol calls. This means your AI assistants can directly connect to this repository, search through your PDFs and Word docs, and pull up specific information on demand. It turns a static file storage system into a living, queryable brain for your AI tools. If you want your agents to know exactly what you know, this is the bridge you need.

🆔 @hackernewsgithubprojects

📦 riba2534/claude-opus-5-5-demo

Claude Opus 5.5 Built Three 3D Games From One Sentence Each

Claude Opus 5.5 Demo is the code generation stress test that proves an AI can build playable, polished 3D games from just a single sentence of instruction. The creator challenged the model to build three distinct web experiences, a pelican riding a bike, a classic first-person shooter map, and a racing game, with zero human edits to the source code. The model handled everything, from researching map layouts to writing the graphics engine and even deploying the live sites itself. This isn't just a demo; it's proof that AI can now handle complex, end-to-end creative tasks without being held by hand, marking a massive leap in autonomous software development.

🆔 @hackernewsgithubprojects

📦 rizzo-ai-academy/rizzo-flow

Rizzo Flow: AI Decisions Without Text

Rizzo Flow is the local AI tool that returns typed decisions without writing a single word of text. Usually, large language models generate slow sentences that you have to parse, but this project reads the hidden probabilities directly from the model’s internal math. It turns messy inputs into instant, structured answers like a clear yes or no, or a specific score. This means you get the judgment of an AI without the lag or format errors of standard text output. It runs entirely on your own machine, making it perfect for private, real-time automation.

🆔 @hackernewsgithubprojects

📦 saragordic/rooms

Rooms

Rooms is the menu bar app that finally tames your chaotic Mac desktop by turning your projects into dedicated virtual workspaces. Instead of drowning in a sea of overlapping windows, you just press option space, type a project name, and your custom layout instantly snaps into place. It hides everything else without actually closing a single app, and it even remembers different layouts for your laptop screen versus your external monitor. When you switch projects, your previous windows are safely parked in the background, ready to return exactly how you left them. It is a brilliant, completely offline way to stay focused.

🆔 @hackernewsgithubprojects

📦 kloxeld/xscrape

xscrape: The Smart Twitter Data Collector

Xscrape is the Python library that finally makes collecting Twitter data feel effortless. It solves the headache of getting blocked by automatically rotating through your account pool and adjusting its speed to stay under the radar. Instead of writing fragile scraping code from scratch, you just tell it what to look for, and it handles the complex background work of managing connections and organizing your results. The coolest part is how it reconstructs entire conversation threads into neat, exportable files, turning messy social feeds into structured data you can actually analyze.

🆔 @hackernewsgithubprojects

📦 v-modal/awesome-jev-tools

One Model for Every Decision

A browser extension uses Jev to detect sponsor reads in YouTube videos with just one dollar per thousand videos. It skips generating text and instead asks a direct question: is this ad or content? The tool returns a simple yes, no, or score, making it fast and cheap for live software. A repository called awesome Jev tools collects these real projects to show how this model works in the wild. It tracks live deployments where Jev handles routing, safety checks, and scoring. This is not a chatbot; it is a decision engine.

🆔 @hackernewsgithubprojects

📦 tianyucodings/jevharness

JevHarness: The AI That Plays Pokemon Better Than You

Take a slow, thoughtful AI and a fast, instinctive one, and let the smart one build a custom program that uses the quick one to win. That is the core trick behind JevHarness. It solves the problem of high cost and slow speed in artificial intelligence. The project uses a large model to write specific code for a task, like playing a battle. It then freezes that code, so a smaller, cheaper model can make rapid decisions without thinking deeply every time. The result is an agent that acts with the speed of a reflex and the strategy of a plan.

🆔 @hackernewsgithubprojects

📦 tencentarc/gamehorizon

GameHorizon: Measuring AI in AAA Games

Tencent ARC just built a benchmark that grades AI models on how well they actually play AAA games, not just on static images. The project is called GameHorizon, and it solves a big problem: figuring out if a model can truly understand complex, long-term strategy in real-time gameplay. It uses five thousand hours of actual game footage to test if AI can plan ahead or just react to what's on screen right now. The results are already out, and the best model barely scored eighty percent. So if you're building agents that need to think several steps ahead, this is the yardstick you need to know where the line is.

🆔 @hackernewsgithubprojects

📦 yinsongxu/llm2jev

How LLM2Jev Answers Questions Instantly

Take a standard local language model and turn it into a rapid decision engine that answers complex questions without writing a single new word. The project called LLM2Jev reads the initial signal from your model to pick an answer or score an image, completely skipping the slow step of generating text token by token. This makes it incredibly fast for structured tasks like choosing between options or evaluating a visual scene. By relying on this prefill-only method, you get precise results in milliseconds rather than seconds. It is a clever way to get smart, structured answers from local hardware without the usual wait.

🆔 @hackernewsgithubprojects