📦 chojs23/ec

Master Your Git Conflicts With This Terminal Tool

Struggling with messy Git merge conflicts in your terminal? Say hello to this powerful, lightweight solution designed for developers who never want to leave their command line. It provides a sleek three-way text user interface that lets you compare yours, theirs, and the base versions side-by-side. With its intuitive, vim-like keyboard shortcuts, you can navigate through conflicts, pick the right changes, and resolve them instantly. It even includes a dedicated Neovim plugin for seamless integration into your existing workflow. Give your productivity a serious boost and make your next code merge a breeze with this incredible developer tool.

📰 https://news.ycombinator.com/item?id=47660167

🆔 @hackernewsgithubprojects

📦 david-oconnor/molchanica

Visualize and Simulate Molecules with Molchanica

Ever wanted to see proteins and molecules in motion right on your desktop? Molchanica is a powerful open source tool that lets you view, edit, and simulate molecular structures with incredible ease. This software bridges the gap between complex research and a smooth user experience, allowing you to run integrated molecular dynamics simulations using Amber force fields or visualize electron density data from crystallography. Whether you are building small molecules in 3D or screening compounds for drug discovery, this platform provides the high-performance visualization you need. Download it today to start exploring the microscopic world with professional precision and real-time feedback.

📰 https://news.ycombinator.com/item?id=47662280

🆔 @hackernewsgithubprojects

📦 kevinrgu/autoagent

Automate Your AI Agent Engineering

Imagine if your AI agent could engineer its own improvements while you sleep. This is what AutoAgent achieves. Instead of manually tweaking your codebase, you define your objectives in a simple markdown file, and let a meta-agent handle the rest. It automatically modifies your system prompts, tools, and orchestration logic, then runs benchmarks to verify performance. If the score improves, the change is saved; if not, it discards it and tries again. It is a brilliant way to evolve your agent architecture autonomously. Start leveraging this automated approach to harness engineering and watch your agents get smarter.

📰 https://news.ycombinator.com/item?id=47662518

🆔 @hackernewsgithubprojects

📦 andmarti1424/sc-im

Ditch Heavy Spreadsheets: Master Your Data in the Terminal with sc-im

Tired of heavy spreadsheet software slowing down your workflow? Meet the Spreadsheet Calculator Improvised, or simply sc-im. This is a powerful, ncurses-based spreadsheet program designed specifically for your terminal. It brings the efficiency of Vim-like keybindings directly to your data management, allowing you to navigate and edit cells with incredible speed. With features like support for massive grids, Lua scripting for custom automation, and direct integration with GNUPlot for visualization, it is a truly versatile tool for power users. If you want a lightweight, keyboard-driven alternative that keeps you productive in the terminal, give this project a try today.

📰 https://news.ycombinator.com/item?id=47662658

🆔 @hackernewsgithubprojects

📦 bilalimamoglu/sift

Stop Letting Coding Agents Get Lost in Noise

Are your coding agents wasting precious time and tokens reading massive, noisy logs? Meet this clever tool designed to turn overwhelming command output into a short, actionable summary. Instead of flooding your AI with raw data, it uses smart, local heuristics to identify root causes and suggest the next logical step before your agent even processes the full file. Whether you are running test suites or parsing complex build errors, this project keeps your workflow fast and efficient by focusing only on the signal. Give your agents exactly what they need to get the job done right.

📰 https://news.ycombinator.com/item?id=47473782

🆔 @hackernewsgithubprojects

📦 fikrikarim/parlor

Real-Time AI Conversations Right on Your Computer

Want to have a face-to-face conversation with an AI that lives entirely on your own machine? Meet this incredible project that brings real-time multimodal interaction to your desktop without needing any cloud servers. By leveraging the power of Gemma 4 for vision and speech understanding combined with the speedy Kokoro engine for voice synthesis, you can talk and show your camera to the AI in real time. It even supports natural features like barge-in, allowing you to interrupt the AI mid-sentence. This is a brilliant way to explore private, offline AI experiences that run locally on your hardware.

📰 https://news.ycombinator.com/item?id=47652007

🆔 @hackernewsgithubprojects

📦 earendil-works/absurd

Build Durable Workflows With Just Postgres

Tired of over-engineering your backend infrastructure? Check out this project which offers a surprisingly simple approach to durable execution. Instead of managing complex external services, this tool runs entirely on Postgres. By decomposing tasks into sequential steps and using database-backed checkpoints, it ensures your long-running functions can survive crashes and network failures without losing any progress or duplicating work. You can manage your queues and inspect task states through its command line utility and web interface. It is a fantastic experiment for anyone wanting to build reliable, fault-tolerant systems using only the tools they already have.

📰 https://news.ycombinator.com/item?id=47665221

🆔 @hackernewsgithubprojects

📦 miracle-wm-org/miracle-wm

Take Control of Your Desktop with Miracle-WM

Tired of rigid desktop layouts that slow you down. Meet Miracle-WM, a highly hackable and modern Wayland compositor that brings professional efficiency to your daily workflow. It solves the frustration of restrictive window management by offering a keyboard-driven experience with smooth animations and versatile tiling options. What makes it truly special is its unique sandboxed WebAssembly plugin system, allowing you to extend core features or write custom window logic in Rust without touching complex C++ code. If you want a stylish, stable, and deeply customizable environment, give this project a look and transform how you interact with your machine.

📰 https://news.ycombinator.com/item?id=47665403

🆔 @hackernewsgithubprojects

📦 panzer-jack/easy-live2d

Add Live2D to Your Web Projects Easily!

Want to bring your Live2D models to life on the web without the usual headaches? This project is an absolute game changer. By building a lightweight wrapper around Pixi.js, it lets you handle complex Live2D animations as simply as moving a basic sprite. It solves the massive complexity of the official SDK, making features like triggering motions, playing voice files, or managing expressions incredibly straightforward. Whether you are using React or Vue, this library streamlines your workflow significantly. Dive into this repository to start animating your characters with ease and take your web projects to the next level today.

📰 https://news.ycombinator.com/item?id=47639734

🆔 @hackernewsgithubprojects

📦 matthartman/ghost-pepper

Turn Your Voice Into Perfect Text Locally on macOS

Tired of your private voice data being sent to the cloud? Meet Ghost Pepper, a powerful macOS tool that brings professional speech-to-text directly to your desktop. Simply hold the control key to record, and when you release, it instantly transcribes and pastes your speech into any application. It runs entirely on your hardware using WhisperKit for accurate transcription and a local language model to automatically clean up filler words and grammatical slips. Because everything happens offline, your sensitive data never leaves your computer. It is the perfect private workflow for fast, polished writing without relying on any external web services.

📰 https://news.ycombinator.com/item?id=47666024

🆔 @hackernewsgithubprojects

📦 pangon/ai-sdlc-scaffold

Master AI-First Development With This Smart Project Template

Stop struggling with fragmented AI workflows and start shipping faster with a truly structured approach. This repository offers a powerful template designed specifically for AI-first software development. By organizing your project into four distinct phases—objectives, design, code, and deployment—it ensures that every decision is captured and documented right within the repository itself. With built-in automation skills for tasks like requirements elicitation and component decomposition, it keeps your context clean and your AI agents highly effective. Use this framework to maintain consistency and traceability from your initial idea to the final product. Give your development process a professional edge today.

📰 https://news.ycombinator.com/item?id=47466513

🆔 @hackernewsgithubprojects

📦 juyterman1000/entroly

Stop AI Hallucinations: Give Your Coding Agent Total Codebase Visibility

Is your AI coding assistant constantly missing important details or hallucinating because it only sees a fraction of your project? Most tools struggle with limited context, but Entroly solves this by acting as a powerful context engineering engine. It compresses your entire codebase into the token budget at variable resolution, effectively stripping out duplicates and boilerplate while maintaining full project visibility. It even uses reinforcement learning to improve its selection over time. Whether you use Cursor, Claude Code, or Copilot, installing this tool helps your AI provide significantly more accurate responses while drastically reducing your token usage.

📰 https://news.ycombinator.com/item?id=47597053

🆔 @hackernewsgithubprojects

📦 anzellai/sky

Meet Sky: The Elm-Inspired Language That Compiles to Go

Ever wanted the functional elegance of Elm paired with the power of Go? Meet Sky, an experimental programming language designed to give you the best of both worlds. It features Hindley-Milner type inference and a unique server-driven UI framework that ditches complex frontend pipelines for seamless, performant web development. Because it compiles directly to a single, portable Go binary, you can ship your entire fullstack application with ease. Whether you are building complex backend systems or interactive interfaces, Sky helps you write safer code faster, leaving the dependency hell behind. Give your development workflow a major upgrade today.

📰 https://news.ycombinator.com/item?id=47662116

🆔 @hackernewsgithubprojects

📦 safishamsi/graphify

Turn Your Messy Codebase into a Smart Knowledge Graph

Ever feel lost in a massive folder of code, PDFs, and random notes? Meet the ultimate AI coding assistant skill that transforms any directory into a structured, queryable knowledge graph. By using deterministic analysis for code and intelligent agents for documents and images, this tool maps out connections, surfacing surprising insights and identifying key concepts that were previously buried. It drastically reduces token usage by letting your AI read a compact map rather than raw files. Whether you are debugging complex architectural decisions or organizing research papers, this utility makes your entire project searchable, understandable, and perfectly connected.

📰 https://news.ycombinator.com/item?id=47668188

🆔 @hackernewsgithubprojects

📦 zoriya/kyoo

Stop Managing Your Media Files Manually With Kyoo

Are you tired of constantly renaming files and organizing folders just to get your media server to recognize your collection. Meet Kyoo, a smart, self-hosted media server designed to handle your movies, series, and anime without the maintenance headache. Unlike other platforms that demand strict folder structures, Kyoo is built to work seamlessly with your existing download directories, even if the file names are messy or unconventional. With powerful features like dynamic transcoding for instant quality adjustments and handy video preview thumbnails, it keeps streaming effortless. If you want a truly set-it-and-forget-it media library solution, give Kyoo a try.

📰 https://news.ycombinator.com/item?id=47668295

🆔 @hackernewsgithubprojects

📦 python-roborock/local_roborock_server

Take Full Control of Your Roborock Vacuum Locally

Ever wish you could manage your smart home devices without relying on external clouds? This project offers a private Roborock server stack that you can host entirely on your own local network. By setting up this system, you gain control over the HTTPS and MQTT communications that your vacuum uses, effectively cutting the cord to external servers. It is a perfect solution for privacy-focused users who want to keep their home data local. Just remember that this is an early beta project, so it is best suited for those comfortable with Docker and technical setups to get everything running smoothly.

📰 https://news.ycombinator.com/item?id=47668576

🆔 @hackernewsgithubprojects

📦 celestoai/smolvm

Build Safer AI Agents with SmolVM

Are you tired of worrying about AI agents wreaking havoc on your local machine? Meet SmolVM, an open-source tool that creates secure, disposable sandboxes for your code execution and browser tasks. It solves the risk of running untrusted code by isolating everything inside lightweight virtual machines, like Firecracker or QEMU. Whether you need to automate browser sessions or provide your agents with a safe shell environment, SmolVM offers a simple Python SDK and CLI to handle it all. It is the perfect way to keep your main system clean and protected. Take control of your agent workflows today.

📰 https://news.ycombinator.com/item?id=47668778

🆔 @hackernewsgithubprojects

📦 mohshomis/modo

Stop Coding Blindly: Meet the Spec-First AI IDE

Tired of AI coding tools that dive straight into code without a plan? Meet Modo, an open-source IDE built for developers who want more structure in their workflow. Instead of jumping from prompt to code, Modo forces a disciplined, spec-driven approach by generating requirements, design documents, and detailed task lists as markdown files before writing a single line of code. You can even use features like agent hooks to automate tasks or steering files to enforce project rules automatically. It is fully open-source, giving you total control and transparency over your development environment. Give Modo a try today.

📰 https://news.ycombinator.com/item?id=47655268

🆔 @hackernewsgithubprojects

📦 arman-bd/guppylm

Build Your Own Tiny Fish AI: Meet GuppyLM

Ever wondered how a massive language model really works under the hood without all the complex technical black magic? Meet GuppyLM, a tiny, specialized language model with just nine million parameters that is specifically designed to talk like a small fish. This project solves the mystery of AI by stripping away the noise to show you exactly how data, tokenizers, and architectures fit together. It is built using a simple vanilla transformer, trained on sixty thousand synthetic conversations in just five minutes. Whether you are curious about artificial intelligence or learning machine learning, dive in and explore today.

📰 https://news.ycombinator.com/item?id=47655408

🆔 @hackernewsgithubprojects

📦 kessler/gemma-gem

Run AI Privately in Your Browser

Ever wanted to run powerful artificial intelligence directly inside your web browser without relying on external servers? The Gemma Gem project makes this possible by utilizing WebGPU technology to execute Google’s Gemma 4 model entirely on your local machine. Because everything happens on your computer, your data never leaves your device and you do not need any complicated API keys or cloud subscriptions. It is a fantastic way to experiment with advanced language models privately and efficiently. Give it a try if you want to keep your digital interactions secure while exploring the latest in browser based machine learning.

📰 https://news.ycombinator.com/item?id=47655367

🆔 @hackernewsgithubprojects

📦 asamassekou10/ship-safe

Is Your AI Code Agent Putting Your Project at Risk?

Are you building with AI agents but worried about hidden vulnerabilities? Modern coding tools bring incredible speed, but they also introduce complex risks like prompt injection, unauthorized tool access, and insecure pipeline configurations. That is where Ship Safe steps in as a powerful security scanner built specifically for the agentic era. By running eighteen specialized agents in parallel, it identifies critical misconfigurations, detects hardcoded secrets, and even vets MCP tool manifests before they connect to your environment. Protect your codebase from supply chain attacks and AI agent misuse today, because staying secure is the smartest way to keep building fast.

📰 https://news.ycombinator.com/item?id=47669335

🆔 @hackernewsgithubprojects