📦 chetaslua/jevmeter

Jevmeter: The BS Meter for Any Video

Jevmeter is the video tool that turns every sentence into a live scorecard. It uses an AI model to analyze your footage and render a cool, sixteen-by-nine edit that flags evasive answers and hype in real time. The best part is that it costs just five cents to process an entire debate. You drop in a file, pick a preset like debate or sales pitch, and it handles the rest. It is the perfect way to turn any talk into a shareable highlight reel. Give it a spin and see who dodges the most questions.

🆔 @hackernewsgithubprojects

📦 tamaratran/fast-jev-compaction

How fast-jev-compaction saves your code context

Shrink your coding bot’s memory without losing the details that matter. Fast Jev Compaction fixes the annoying habit of AI tools forgetting file paths or exact errors just to fit more text into the window. Instead of summarizing your chat, it asks a second model to look at your entire history and decide exactly which old tool results to drop. Anything it keeps stays word for word, so you never lose a critical command. This plugin turns messy, bloated conversations into a clean, precise record. Your bot remembers the important stuff, not just a vague summary. It’s a simple swap that keeps your coding sessions sharp and accurate.

🆔 @hackernewsgithubprojects

📦 arahim3/cachebeat

Cachebeat: Stop Paying Full Price

Cachebeat is the tiny Claude Code skill that stops your conversation from paying full price just because you stepped away. It solves a sneaky billing trap where long AI sessions lose their memory discount after an hour of silence, causing the entire chat history to be re-processed at ten times the normal cost. The surprising part is that it uses a simple inactivity timer to send a few words of text, like saying "still here," which resets the clock and keeps the cache warm. This means your next message reads from the cheap cache instead of triggering an expensive rebuild, saving you real money or protecting your subscription quota.

🆔 @hackernewsgithubprojects

📦 magiccodingman/vllm-radiance

VLLM Radiance: The AMD GPU Inference Secret

This project runs large language models on AMD Radeon AI Pro R9700 graphics cards by compiling a custom version of the VLLM inference engine specifically for that hardware. Most standard AI servers are built for Nvidia chips, so running complex models on AMD cards is usually slow and unstable. This repository fixes that by using hand-written math routines that match the specific architecture of the R9700 GPU, allowing it to serve text and vision data at high speeds. It is an experimental but highly tuned setup that proves you can get serious performance out of consumer-grade AMD silicon.

🆔 @hackernewsgithubprojects

📦 scrypthunter/comfyui-yue2

ComfyUI Yue2: The Safe Song Maker

ComfyUI Yue2 is the music node that finally lets you generate full stereo songs without crashing your setup. It solves a nasty problem where installing new music tools would break your existing image generation software. The fork loads models safely in isolation, so you can turn text and lyrics into a complete track or extract a score from a reference song to create a cover. It fixes the dependencies that usually cause conflicts, meaning your whole system stays stable while you create music. If you want to make songs on your computer without losing your image workflows, this is the one to try.

🆔 @hackernewsgithubprojects

📦 theoleecj/semif

SemIf: AI Decisions Without the Fluff

SemIf is the open source tool that reads AI decisions directly without waiting for text to generate. Usually, when you ask an AI to make a simple choice, like routing a ticket or verifying a fact, it has to write out a full sentence before your software can actually understand the answer. It is slow and wasteful. SemIf skips that entirely. Instead of forcing the model to speak, it peeks directly at the raw numbers inside the model to see which option is most likely. This means you get the answer in about one second instead of five.

🆔 @hackernewsgithubprojects

📦 maksshandybo/perestruct

PereStruct: The Soviet Newspaper Solver

PereStruct is the document pipeline that finally makes sense of messy Soviet newspaper scans. It takes a chaotic page of text, uses computer vision to find where the articles start and stop, and then uses language models to clean up the errors. The coolest part? It figures out which scattered blocks belong to the same story, even if they are on different columns. It actually beats out massive vision-language models at this task. So if you have a pile of old, hard-to-read news clippings, this is your secret weapon for turning them into clean, readable text.

🆔 @hackernewsgithubprojects

📦 opendfm/image_edit_agent

Chat to edit photos like a pro

Fix your photos just by talking. A new research project called Image Edit Agent turns casual speech into precise photo adjustments, letting you tweak exposure or contrast without touching sliders. Instead of generating random images, it controls standard editing tools so you can see exactly what changed and fix mistakes easily. This makes professional retouching accessible to anyone, turning a complex task into a simple conversation. It is a practical step toward truly understandable and controllable AI image editing.

🆔 @hackernewsgithubprojects

📦 jarrodwatts/jev-trader

Jev Trader: AI Trading at 300 Milliseconds

Place a real trade on the blockchain every three hundred milliseconds. Jev Trader uses an AI model called Jev to watch the order book for the Monad and USDC pair. Every single block, the AI makes a split-second decision to buy or sell. It then sends a live order from a real wallet, landing in the same block it made the choice. The whole process, from reading the market to posting the trade, fits into a tight 300-millisecond window. A dashboard streams every decision, order, and fill live, so you can watch the AI trade in real time.

🆔 @hackernewsgithubprojects

📦 fhshaik/typesafe-mario

How an AI Plays Mario

TypeSafe Mario lets a language model play Super Mario Bros without looking at the screen. Instead of processing visual pixels, the system reads the game’s internal memory and converts it into a compact text description of the world. It tells the model exactly where Mario is standing, how fast he is moving, and which enemies are approaching. The model then chooses one simple physical action, like jumping or running right. This approach is interesting because it shows how far we can push AI reasoning by giving it clean, structured data rather than raw images. It strips away visual noise to focus purely on logic and timing.

🆔 @hackernewsgithubprojects

📦 usenotra/notra

Notra: Turn Your Work Into Content

Notra scans major artificial intelligence engines to see exactly where your brand appears in answers and who gets recommended instead. It tracks specific buyer questions across platforms like ChatGPT and Gemini to measure your visibility and share of voice. The platform identifies gaps where you are missing and helps you turn those opportunities into published content drafts. It also attributes traffic to distinguish genuine human referrals from simple crawler visits. This tool solves the problem of being invisible in AI-driven search by providing clear data on your standing. It allows teams to fix their digital presence with actionable insights.

🆔 @hackernewsgithubprojects

📦 realzachi/pg-jev

pg-jev: Ask your Postgres tables questions in plain language

You can now filter your Postgres database using plain English instead of complex SQL logic. The project called pg-jev lets you ask questions like the name is European right in your where clause. It solves the problem of writing brittle rules for messy human data by using an AI model that returns a simple yes or no with a confidence score. No vector databases or embeddings are needed here. It is clever because it treats the AI judgment as a standard SQL function that works perfectly with your existing joins and limits. If you hate writing fuzzy search logic, this is the tool you need to know about.

🆔 @hackernewsgithubprojects

📦 kitze/skillbox

skillbox: The Library for Your AI Agents

Skillbox turns your local AI agent skills into a proper versioned library that lives entirely on your own server. The surprising part? It acts like a strict gatekeeper for your agents, letting you track every change, roll back mistakes, and hand out specific keys so different bots only see the exact instructions you allow. It plugs into standard tools through a dedicated bridge, meaning your agents can search, load, and verify their capabilities without ever touching the internet or a third-party cloud. You own the data, you control the access, and the whole thing runs right on your machine.

🆔 @hackernewsgithubprojects

📦 thruwire/foreman

Foreman: The AI Supervisor for Coding Agents

Foreman is the supervisory layer that watches AI coding agents while they work. Instead of just letting a slow language model code, Foreman adds a fast, independent brain that constantly asks nine specific questions about progress, stuck states, and quality. It runs in parallel with the coder, so the agent never has to stop to be checked. If the AI gets stuck or drifts off track, Foreman can gently steer it back or step in. It is a clever architectural experiment that separates the heavy thinking of coding from the light thinking of supervision. The result is a smarter, safer way to manage autonomous software projects without constant human babysitting.

🆔 @hackernewsgithubprojects

📦 devagrawal09/jev-review

Jev Review: The Code Review That Actually Thinks

Jev Review is the code-review tool that stops guessing and starts proving. It scans your entire codebase, scores every file for specific risks, and only digs deeper when the numbers look suspicious. The coolest part is the dashboard: it shows you exactly which files tripped the alarm and why, turning a chaotic sea of code into a clear, prioritized list of actual problems. It is not a generic linter; it is a structured investigation into what might break. If you want to know where your code is actually weak before it goes live, this is the shortcut you have been missing.

🆔 @hackernewsgithubprojects

📦 mcncarl/jianying-headless

Jianying Headless: Automating Video Editing

Your computer can now edit professional videos without ever opening the editing app, and that is exactly what this project does. Called Jianying Headless, it lets you write a simple plan describing your cuts, text, and audio, then automatically builds a fully editable project file that the desktop editor opens natively. It works by quietly talking to the installed app in the background to assemble the timeline and export the final video. No mouse clicks, no manual dragging, just code. It is a fascinating look at how far automation can take creative tools, turning a complex graphical interface into something you can script and repeat.

🆔 @hackernewsgithubprojects

📦 droidrun/mobile-jev

Mobile Jev: The Phone Agent That Actually Works

Mobile Jev is the mobile agent that makes real decisions on a live phone without needing a direct cable connection. It solves the headache of writing brittle scripts for every app change by letting an AI model look at the screen and choose the right action on its own. The coolest part is that it does not just claim success; it actually checks the phone to confirm the job is done. You can watch it open apps and tap buttons in a live web studio, giving you a true view of what the agent is doing. It is a slick way to see how far autonomous phone control has come.

🆔 @hackernewsgithubprojects

📦 danjdewhurst/story-skills

Story Skills: The Continuity Engine

Story Skills catches dead characters walking, unfired Chekhov guns, and payoffs that land before their setup before a reader ever could. It’s a set of agent skills that gives AI writing tools a shared, plain-markdown format for your entire story bible, including character files, worldbuilding notes, and plot arcs. But the real magic is the companion command line tool that treats your story like code. It runs a deterministic continuity engine that checks for contradictions exactly like a compiler treats type errors. If your files are well-formed but the plot doesn’t hold together, it flags the exact file and line.

🆔 @hackernewsgithubprojects

📦 gentleman-programming/gentle-shell

Gentle Shell: The Architect's Terminal

gentle shell turns your terminal into a controlled workspace where you direct your coding agent instead of just typing commands. It solves the problem of messy, unstructured AI coding by giving you a single place to see active tasks, track changes, and review evidence. The most interesting part is how it acts like a senior architect; it doesn't just write code, it plans, verifies, and shows you exactly what it did so you can trust the result. If you want your AI helper to feel like a reliable colleague rather than a random text generator, this is the tool that makes that possible.

🆔 @hackernewsgithubprojects

📦 local-inference-lab/b12x

Run DeepSeek on Your RTX 5090 with B12X

Run DeepSeek V4.1 on your RTX 5090 or DGX Spark. B12X is a kernel library that turns consumer Blackwell cards into a local inference machine. It handles the low-level math for attention and experts, letting you serve large models right on your desktop. No datacenter needed. If you want to run serious AI on your own hardware, this is your shortcut. Grab B12X and start experimenting with local models today.

🆔 @hackernewsgithubprojects