📦 tencentarc/player-1

Player-1: Teaching AI to Play from Synthetic Experience

Player-1 teaches AI models to control video games by letting them practice on synthetic experience rather than watching humans play. The core idea is simple: instead of relying solely on human demonstrations, the system first fine-tunes a computer-use model on curated gameplay data, then sends it into a gym of sixteen synthetic environments where it interacts with verifiable rewards to build real experience. This approach helps the agent learn general game control—aiming, driving, navigating, and using interfaces—across diverse 2D and 3D scenarios.

🆔 @hackernewsgithubprojects

📦 seerray-lab/xiaomi-ocr-0

Xiaomi OCR Zero: Tiny Model, Big Reads

Xiaomi OCR Zero is the tiny language model that finally reads your messy papers with a giant brain. It is a small, efficient engine that turns images of receipts, handwritten notes, and dense math equations into clean, usable text without needing a cloud server. The real surprise is that this compact model matches the performance of much larger systems, handling everything from ancient calligraphy to modern spreadsheets with high accuracy. You get a local tool that understands what you see, not just what is written. It is a powerful proof that you do not need massive hardware to get professional-grade reading results on your own machine.

🆔 @hackernewsgithubprojects

📦 phuryn/work-humanizer

Work Humanizer: The AI Tone Remover

Work Humanizer is a specialized tool that rewrites professional text to sound like a person instead of a robot. The most surprising part is not just the rewrite, but the rigorous proof that it works. The creator used blind judges and hand-checked every single output to guarantee that the meaning stayed exactly the same while the robotic flavor vanished. This project solves a very real problem. When you write with AI assistance, your words can start to feel generic and cold. This tool strips away that artificial feel so your emails, posts, and memos sound like you.

🆔 @hackernewsgithubprojects

📦 jlevy/tbd

tbd: The Git-Native Task Tracker for AI Agents

tbd is the Git-native task tracker that finally gives your AI coding agents a shared, persistent memory. Instead of losing context between chat sessions, this tool stores your tasks, bugs, and plans directly inside your repository as simple Markdown files. That means any agent on any machine can pick up exactly where you left off, keeping your team perfectly in sync without needing a heavy external database or cloud service. It is the missing piece of infrastructure that makes your automated coding workflows feel genuinely collaborative, letting you plan, review, and ship code with total confidence.

🆔 @hackernewsgithubprojects

📦 coding-by-feng/ai-agent-session-center

See Every AI Agent at Once

Spot the exact AI coding agent waiting for your approval. Agent Session Center turns a chaotic jumble of terminal windows into one clear dashboard. It watches your Claude Code and Codex sessions, letting you see exactly who is stuck, finished, or running. You can queue the next prompt right there, so you stop tab-juggling and only step in when it truly matters. The app even remembers your workspace, so you can pick up right where you left off. It is the simplest way to manage multiple AI helpers without losing track. Check out the repository to keep your coding agents in check.

🆔 @hackernewsgithubprojects

📦 atoll92/agent-standup

Agent Standup: Your AI Team in a Live Call

Turn your coding assistant sessions into a live, animated team meeting where you can actually see the work happen. This tool reads the files your AI already writes and turns them into a video call with voices, expressions, and screen shares. You see exactly who is working, who is stuck, and what they are changing in real time. It is a simple way to monitor complex AI tasks without staring at terminal logs all day. Watch the agents talk, type, and share their screens as if they were sitting in a room with you. It is a clear, visual way to keep an eye on your digital team.

🆔 @hackernewsgithubprojects

📦 cosmicfrontierlabs/unlinked

Unlinked: Run Simulink in the Browser

Open your Simulink models directly in the browser without installing a single thing. Unlinked is a web app built in Rust that lets you view, edit, and run simulations right from your desktop. It even takes your MATLAB code and turns it into standalone Rust code you can use anywhere. You can drag blocks around, tweak parameters, and watch the results stream in live. It’s not a full replacement for the whole suite, but for the parts it supports, it’s fast, private, and surprisingly capable.

🆔 @hackernewsgithubprojects

📦 andrei-voia/tilemakerdot

TileMaker DOT: The Pixel-Smart Map Editor

TileMaker DOT automatically figures out which parts of your game characters should appear behind or in front of other objects by analyzing the actual pixels of your images. This removes the tedious need to manually manage layers in complex software, ensuring your 2D game maps look right instantly. This Java-based editor is built specifically for 2D game developers who want to focus on design rather than fighting their tools. It handles everything from importing spritesheets to exporting maps for engines like Unity and Godot, all while keeping your assets perfectly organized.

🆔 @hackernewsgithubprojects

📦 mcg-nju/videojev

VideoJev: The Fast Video Model

VideoJev is the video model that answers questions without writing a single word. Most systems guess answers by generating text, which is slow and prone to errors. VideoJev skips the writing entirely. It scores the likelihood of every possible answer in one go. This means it finds the exact moment a person eats pasta or picks an object much faster than traditional tools. It turns video understanding into a simple scoring game rather than a creative writing exercise. This approach makes real-time interaction possible. If you want to build apps that react to video instantly, this is the tool that changes the speed.

🆔 @hackernewsgithubprojects

📦 gongnyang/awesome-html-scrolline-deck

Scrolline: Scroll-Driven Presentation Decks

A single bar in a chart physically travels across the screen to become a label in the next scene, and the audience follows that object rather than watching a slide cut. This is the core idea behind a project called awesome html scrolline deck. It builds presentation decks as one continuous timeline driven by scrolling, so every transition is a handoff where an element from one beat moves into its place in the next. The result is a smooth, narrative flow that feels like a film rather than a slideshow.

🆔 @hackernewsgithubprojects

📦 ashishkumar4/armada

How to Run Code on a Hundred Cloud Containers in Seconds

Run your code on a hundred cloud containers at once to finish heavy jobs in seconds. That is exactly what armada does. It takes a single command or function and spreads it across many small machines in your own cloud account. The most impressive part is its speed. A test suite that normally takes thirteen minutes on standard continuous integration tools finished in just seven and a half minutes on armada. It also re-encoded a hundred videos in twenty-two seconds. Instead of waiting for one slow machine, you get results as they land. It handles the setup, the scheduling, and the cleanup automatically.

🆔 @hackernewsgithubprojects

📦 arahim3/digup

DigUp: Search Your Mac By Meaning

DigUp lets you hunt through your Mac by describing what is actually in your files, rather than guessing filenames. It uses an AI model to look inside photos, listen to podcasts, and scan documents so you can find that one specific moment or clause instantly. The coolest part is that it works on your Mac, so your private files never leave your computer. It turns your library into a searchable brain where a picture of a dog matches the words dog on the beach. No more digging through folders for hours. Just describe the scene, and DigUp finds the exact spot for you.

📰 https://news.ycombinator.com/item?id=50035576

🆔 @hackernewsgithubprojects