📦 svy04/ballast

ballast: The memory system for your AI

ballast stops your AI assistant from forgetting what you already told it. It takes a big goal you don't fully understand and breaks it into small, checkable steps. The coolest part? It remembers. Every rule, decision, and fact you pin gets saved in plain text files, so your AI actually carries that weight into the next session. You stop repeating yourself, and your work builds on itself instead of starting from zero every time. It is a simple, clever way to keep your digital brain steady. Try it, and watch your projects get way less chaotic.

🆔 @hackernewsgithubprojects

📦 stackblitz/bolt-slides

Bolt Slides: AI Builds Real Web Apps

Create presentations that are actually working web applications, not static slides, using a single prompt. This project, called Bolt Slides, provides a foundation for AI agents to build interactive decks where every slide is a live, responsive React component. It solves the problem of generic, boring slide outputs by allowing developers to embed real code, live data, and even 3D scenes directly into their pitch decks. Instead of just text and images, you can present actual working prototypes or dynamic charts. The result is a fully functional web app that you can share as a simple link, making your presentations far more engaging and professional.

🆔 @hackernewsgithubprojects

📦 dnv2003/v100-skinny

Running 2025 AI on 2017 GPUs

Make modern AI run fast on old hardware. The project called v100 skinny lets you serve the heavy Qwen model on four used Tesla V100s, which lack native support for the new four-bit precision format. Instead of slow conversion tricks, it uses hand-written code to decode weights directly inside the chip, hitting speeds that rival a brand-new RTX 5090. This proves you can get serious performance from affordable, aging hardware without losing accuracy, turning budget GPUs into serious workhorses for developers who need speed without the latest price tag. If you have older gear sitting idle, this is your proof that software can bridge the gap.

🆔 @hackernewsgithubprojects

📦 m1ckc3s/nullframe

Nullframe: Your Browser's Hidden Instrument Panel

Turn your browser tab into a live instrument panel that actually listens to your computer. Nullframe grabs real data like your frame rate, battery life, and network speed, then renders it in a sleek, industrial grid that feels like a high-end workstation. The coolest part is how honest it is. It clearly tags every widget as either live or simulated, so you never confuse real telemetry with fake numbers. It even beeps softly when you hover, using pure audio code with no libraries. This is a masterclass in clean, efficient web design that runs cool even on your phone.

🆔 @hackernewsgithubprojects

📦 geeklee/srt-whiteboard-animation

SRT Whiteboard Animation

Srt Whiteboard Animation is the subtitle tool that turns plain text captions into hand-drawn ink animations. It takes a standard Srt file and renders it onto a warm, beige paper background, mimicking the look of a classic whiteboard sketch. The project uses a clever masking technique to manage how different parts of the text appear, while the stream feature creates a continuous flow of ink that transitions smoothly into color. This solves the problem of static captions feeling boring by making the text feel alive and drawn in real time.

🆔 @hackernewsgithubprojects

📦 alattanzio/box3dunreal

Box3D for Unreal: The Physics Swap

Box3DUnreal is the Unreal Engine plugin that lets you swap out the default physics engine for a lighter, faster alternative without rebuilding your entire project. It brings Erin Catto’s Box3D physics into Unreal Engine five so you can test custom gameplay systems that need precise, deterministic movement. The cool part is that it runs alongside your existing setup without breaking anything, letting specific objects use this new physics layer while the rest of the world stays on the default engine. It is a clean, lightweight way to experiment with how things move and collide in your games, giving you a fresh tool to build unique mechanics without heavy overhead.

🆔 @hackernewsgithubprojects

📦 abbosaliboev/free-ai-bible

Free AI Bible

Groq offers fourteen thousand free LLM requests every single day without asking for a credit card, which is the core insight behind a new resource called Free AI Bible. This repository is a massive, hand-tested directory of seven hundred free AI tools, APIs, and hosting platforms that allow developers to build production-grade applications on a strict zero-dollar budget. It solves the problem of inflated cloud computing costs by mapping out exactly where to find free LLMs, image generators, and local model runners that genuinely work.

🆔 @hackernewsgithubprojects

📦 drobins25/craft

Craft: The AI Plugin That Remembers Your Taste

Craft is the coding plugin that finally remembers what you love. It works inside your existing Claude Code setup but changes the game by making your codebase read only by default. Instead of just following orders, Craft acts as a creative workshop where specific agents handle research, design, and validation. The coolest part is the mockup funnel. You describe a visual idea, and it builds three live options right in your browser. When you pick the one you like, Craft doesn't just move on. It locks that specific design choice into a permanent standard for your project. From that point on, it enforces your taste.

🆔 @hackernewsgithubprojects

📦 abdomk1998/awesome-digital-escape-tools

Curate Your Own Digital Escape Map

Curate a living directory of privacy tools that help you reclaim your digital life. This repository automatically generates an open source list from the digital escape tools website, gathering open source alternatives for apps you want to replace. Instead of sifting through endless lists, you get a structured, searchable guide that highlights the most essential privacy tools. It works by syncing directly with the live data, ensuring the recommendations are always current and accurate. This is perfect for developers who want a clean, readable format for tracking privacy software. The project bridges the gap between a web directory and a developer friendly format.

📰 https://news.ycombinator.com/item?id=49309959

🆔 @hackernewsgithubprojects

📦 giannisanni/pulsar

Running a 700B Model on Your Laptop GPU

Run massive AI models on hardware that technically lacks the space to hold them. Pulsar solves this by keeping only the active parts of the model in your GPU memory while streaming the rest directly from your fast solid-state drive. It automatically measures your hardware speed to decide which layers stay resident and which stream in, turning your disk into an extended memory bank. This allows you to process huge, complex models with just a few consumer graphics cards and no expensive server setup. The result is a practical way to test giant models locally without waiting for hardware upgrades.

🆔 @hackernewsgithubprojects

📦 alibaba/anolisa

Anolisa Cuts AI Costs by Compressing Data

Alibaba has built Anolisa, an operating system layer specifically designed for AI agents that can drastically reduce your bill by stripping unnecessary data before it reaches the model. It acts like a smart filter for your terminal, keeping your existing setup but allowing the agent to work directly inside it. The standout feature is its token compression, which removes redundant tool outputs to save money without changing how you code. It also uses deep system tracking to show exactly where your computing power is going. If you are running agents, this is how you keep them efficient and cheap while staying in the command line you already know.

🆔 @hackernewsgithubprojects

📦 yjh051108/dsh-router-standard

The Secret to Perfect AI Coding

Dsh Router Standard is the configuration layer that fixes the personality split in your coding assistant. Imagine using the same AI for both building a new app and fixing old bugs. It often acts like two different people, getting one right and the other wrong based on how it starts. This project measures that behavior shift and uses a router to pick the best mode for you automatically. It stops the AI from guessing and forces it to commit to the right approach from the very first second. That means your code actually works the first time.

🆔 @hackernewsgithubprojects

📦 askie/grix

Grix: Your AI Agents as Chat Contacts

Grix turns coding agents into actual chat contacts, letting you text them like you would a colleague. Instead of staring at a terminal, you assign tasks in a group chat and check progress from your phone. It supports fifteen different AI tools, so you can mix and match your favorite assistants into one shared team. You can even pause work or take over control at any moment, keeping everything in one clear thread. It is a smart way to manage digital workers without losing your train of thought. Try moving your code agents into your pocket.

🆔 @hackernewsgithubprojects

📦 shadcnstudio/shadcn-nextjs-admincn-admin-template-free

AdminCN Free: The Ultimate Starter

AdminCN Free is the admin dashboard template that actually saves you weeks of work. Most free templates feel empty, but this one is packed with real pages, including a full email client, calendar, and user management system, all built with the popular Next.js and Tailwind stack. It includes everything you need, from login screens to complex data tables, so you can stop wrestling with basic layouts and start building your actual application logic immediately. It is the perfect head start for developers who want a clean, responsive foundation without the headache of setting up every single component from scratch.

🆔 @hackernewsgithubprojects

📦 matraix-ai/matraix-persona-8b

MatrAIx Persona 8B: Testing Apps with 8 Billion Virtual Humans

Imagine testing your next app not with a generic user, but with millions of distinct digital personalities. The MatrAIx Persona 8B project allows developers to simulate realistic human behavior at a massive scale before launching anything to the real world. By creating detailed persona records that mimic diverse psychology and capabilities, this infrastructure lets you run your software through complex scenarios in chat, web, and native app environments. It turns vague user feedback into concrete, reproducible data. You can stress-test features against a population of simulated users to find edge cases you might have missed.

📰 https://news.ycombinator.com/item?id=49370103

🆔 @hackernewsgithubprojects

📦 indigo0445/yt-collate

The YouTube Music Client That Lives In Your Terminal

Stop opening your browser for music. Collate lets you play, search, and manage your entire YouTube Music library right inside your terminal, using only your keyboard. It is a fast, lightweight tool built for Linux that handles playback, playlist editing, and even downloading tracks without ever leaving your code editor. The coolest part is the Vim-style workflow, where you can copy and paste songs between playlists just like text, making library management surprisingly intuitive. If you love efficiency and hate tab fatigue, this is your new favorite daily driver.

📰 https://news.ycombinator.com/item?id=49356485

🆔 @hackernewsgithubprojects

📦 calnode/calnode

Calnode: The Self-Hosted Scheduling Engine for AI Agents

Calnode is a self-hosted scheduling app that runs entirely from a single static binary, meaning no external database servers are required. It combines built-in video meetings with an AI notetaker, all wrapped in a tiny footprint that fits on a five dollar virtual server. The standout feature is its native support for AI agents, allowing tools like Claude to find open slots and book meetings directly through a standard protocol. This removes the need for manual setup or complex API integrations. The codebase is deliberately small, making it easy to audit and trust.

🆔 @hackernewsgithubprojects

📦 osolmaz/pi-workflows

Watch Your AI Agent's Workflows Run Live

pi-workflows is the workflow engine that lets your coding agent run multi-step tasks you can actually watch in real time. You define these workflows as simple code graphs, and the system executes them directly inside your existing chat session, keeping all the context you already built up. The standout feature is the live terminal viewer, which draws a dynamic map of your workflow, showing exactly which step is running, where it paused, and how it will route next. This solves the problem of opaque automation, giving you a clear visual trace of the logic while it executes.

🆔 @hackernewsgithubprojects

📦 analogy-ai/fm-bench

FM-Bench: The Football Game That Tests AI Patience

FM-Bench is a football management game where the real opponent is your own attention span. You run a club for twenty in-game years, making hundreds of decisions with only noisy scout reports and no visible player stats. It is designed to expose where language models fail over long periods, specifically when they forget their own plans or leak money through sloppy tools. The engine is deterministic, so every run is a perfect audit trail. If your agent can keep a coherent strategy across decades of simulated time without the board firing them, you have found a weakness in current AI.

🆔 @hackernewsgithubprojects

📦 luckygirl-xu/awesome-agent-dynamic-graphs

Your AI Agent is a Living Network

That massive wall of text isn't a coding tutorial, it is a map of how AI agents actually change. Researchers just realized that building smarter AI is like growing a living network, not just writing code. They call this view self evolving agents as dynamic graph transformation. Instead of treating an agent as a static box, we see it as a constantly shifting web of memories, tools, and skills. When an agent learns something new, it doesn't just memorize a fact, it physically restructures its own internal wiring. This repository collects the papers proving that this biological style of growth is the future.

🆔 @hackernewsgithubprojects