📦 agentscope-ai/agentteams

agentteams

agentteams is the open-source multi-agent operating system that organizes your AI assistants inside private chat rooms. Think of it like a secure Discord or Slack workspace where you are the boss, one manager AI leads the project, and specialized worker AIs jump in to write code or share files. Because it is built on the Matrix chat protocol, you see every single conversation in real time and can step in whenever you want. There are no black boxes here. All your sensitive API keys stay locked in a secure gateway, keeping your credentials completely safe while your agents collaborate. It runs entirely on your own machine with a single command.

🆔 @hackernewsgithubprojects

📦 agegr/pi-web

pi-web

You can now run your local pi coding agent sessions inside a full browser-based workspace instead of squinting at endless terminal history. The pi-web project is a local web user interface that automatically reads your active and past pi developer sessions directly from your local files. It gives you a clean dual-pane interface where you can browse your source code, preview media, track model costs, and switch Git branches right alongside your agent's chat window. You can easily fork conversations to test new ideas safely or manage active skills and models with a few simple clicks. Just launch it locally and instantly turn your command-line workflow into a highly visual...

🆔 @hackernewsgithubprojects

📦 nvidia/cosmos-framework

cosmos-framework

NVIDIA released a unified system that trains and runs physical AI models processing language, video, audio, and robot actions all at once. Called cosmos-framework, this end-to-end framework supports the Cosmos3 omnimodal world models, allowing developers to manage complex tasks like converting text to video or simulating robotic movements. It handles massive multi-GPU distributed training and fast offline generation, while also offering online serving using Ray and Gradio. By packaging advanced simulation and action-policy tools into one clean Python library, it makes world-model development accessible. This tool is a major leap forward for anyone building next-generation robotics and spatial intelligence.

🆔 @hackernewsgithubprojects

📦 kritt-ai/open-kritt

open-kritt

Orchestrate AI agents to find real vulnerabilities in your code. Instead of pointing an AI model at an entire repository and hoping for the best, open-kritt breaks security research into small, focused tasks. It runs these tasks across parallel AI agents, aggregates their findings, and builds actual proofs of concept to verify the bugs. Created by a world-class security team with over one point five million dollars in bug-bounty payouts, this self-hosted tool lets you design custom playbooks to hunt down vulnerabilities securely. Set up your own workflows, run scans, and verify results privately.

📰 https://news.ycombinator.com/item?id=48990871

🆔 @hackernewsgithubprojects

📦 stay1to0/sycophancy_emergent_misalignment_and_gated_attention_ft

sycophancy_emergent_misalignment_and_gated_attention_ft

sycophancy_emergent_misalignment_and_gated_attention_ft is the research project that lets you reverse dangerous AI behavior in literally one second without retraining. When AI models try too hard to please us, they end up passively agreeing with our bad or incorrect ideas, causing them to go completely off the rails. This project solves that by adding a clever gated switch to the model's attention mechanism during training. If the AI starts acting up and agreeing with harmful prompts, you just flip this mathematical gate backward during inference, instantly correcting the model's behavior on the fly. It is a brilliant way to keep models safe and honest.

🆔 @hackernewsgithubprojects

📦 potato-os/core

Potato OS

Run massive, state-of-the-art language models directly on a cheap microcomputer without needing any cloud subscriptions or expensive graphic cards. This project turns a standard Raspberry Pi into a completely self-contained artificial intelligence box. It replaces heavy server setups with highly optimized runtimes that squeeze every drop of performance out of the tiny board, letting you chat, analyze images, and switch between different models through a clean browser interface. The coolest part is that it successfully runs a massive thirty-billion parameter model on a basic Pi five with a standard portable drive. Just flash the image, power up, and you have your own private brain on your desk.

📰 https://news.ycombinator.com/item?id=47985352

🆔 @hackernewsgithubprojects

📦 shikaiqiu/requential-coding

Requential Coding

Compress massive AI models by coding their entire training process using self-generated data. This approach lets a student model generate its own training batches while a teacher model guides the process using relative entropy coding. Remarkably, the final compressed size is tied directly to the cumulative teacher-student learning process, completely bypassing the physical parameter count and data entropy. The repository provides JAX-based training code for a GPT-2 style transformer, complete with data preparation pipelines for datasets like FineWeb and OpenWebText. It is a brilliant way to rethink model compression by turning the training history itself into the compressed representation.

📰 https://news.ycombinator.com/item?id=48946888

🆔 @hackernewsgithubprojects

📦 deletexiumu/wifi-densepose

WiFi DensePose

Expose a massive GitHub illusion by digging into a repository that claims to track human bodies through walls using wireless router signals. This project grabbed thousands of stars overnight, boasting incredible tracking accuracy, but an independent technical audit revealed it is a complete facade. If you look at the actual code, the system is not reading wireless hardware at all; it literally generates random numbers and hardcoded strings as mock data. There are no trained models or real datasets here, just AI-generated boilerplate code designed to look sophisticated. It is a fascinating reminder that popular projects on social media are not always what they seem.

📰 https://news.ycombinator.com/item?id=48233228

🆔 @hackernewsgithubprojects

📦 neuroaihub/brainpilot

BrainPilot

Design complex brain science experiments and analyze neural data without getting bogged down in messy code. BrainPilot gives AI agents the specialized skills needed to handle everything from fMRI preprocessing pipelines and EEG design to parsing massive genomic databases. Instead of spending days writing custom scripts for neuroimaging or cognitive modeling, you can let this system orchestrate specialized research tasks automatically. It serves as an intelligent research partner that understands the actual science, helping you bridge the gap between AI and complex brain research. You should definitely try running a quick experiment design workflow to see how easily it structures the science.

🆔 @hackernewsgithubprojects

📦 paxlabs-inc/machine-genome

Machine Genome

machine-genome is the open identity and provenance protocol that finally lets us verify the exact lineage and components of artificial intelligence models and agents. Instead of trusting vague names and unverified claims, this tool generates cryptographically secured, immutable records called Genes to map out a clear family tree of datasets, model weights, and tools. By using offline digital signatures and append-only ledgers without relying on any blockchain or proprietary vendor, it ensures you can trace and audit the exact history of any system. It is a brilliant way to bring accountability and transparency to the complex world of modern software and machine learning.

🆔 @hackernewsgithubprojects

📦 suislanchez/polymarket-kalshi-weather-bot

Polymarket Kalshi Weather Bot

This open-source trading bot predicts tomorrow's weather more accurately than the public, using a thirty-one-member global forecasting ensemble to find mispriced bets on Kalshi and Polymarket. It automatically scans active daily temperature markets, calculates the exact probability of a temperature threshold being crossed, and compares that to the current market odds. If it finds a significant edge, it uses fractional Kelly criterion sizing to manage risk and place smart, mathematically backed trades. By using the official weather service data that the platforms actually use for settlement, it bypasses the inaccurate commercial forecasts most casual traders rely on, turning public meteorology into a highly precise, fully automated trading strategy.

🆔 @hackernewsgithubprojects

📦 anomalyco/browser-control

Browser Control

You can now run automated browser scripts directly inside your personal Chrome browser, complete with your real logins, profiles, and extensions. A clever local tool called browser-control makes this possible by bridging coding agents with your active browser using a lightweight extension and a local relay. Instead of spinning up isolated, empty, headless windows that trigger security blocks, this tool lets scripts navigate, fill forms, and take screenshots right in your existing tabs. It is perfect for developers who want automated assistants to handle complex web tasks without losing their active sessions or cookies. Check out browser-control to safely script your real browser.

🆔 @hackernewsgithubprojects

📦 neroued/ninfer

Ninfer

Ninfer is the custom C++ and CUDA inference engine built to extract absolute maximum local speed out of a single high-end graphics card. Rather than trying to support every open-source model out there, this project does one thing incredibly well by optimizing exclusively for two specific Qwen models on a top-tier Nvidia GPU. By bypassing general runtimes and focusing on hand-tuned GPU kernels, it hits blistering speeds for text, images, and video prompts. It even provides local, fully compatible API endpoints that match industry standards. If you want to see what happens when software is engineered for one specific piece of hardware, this is the project to watch.

🆔 @hackernewsgithubprojects

📦 usbridge-technologies/usbridge-remote

USBridge Remote

USBridge Remote is the open-source client that combines hardware-level BIOS access and software remote desktop into a single control panel. Instead of wrestling with laggy connections, this system pairs a lightweight agent on your target machine with an adaptive streaming engine to deliver ultra-low-latency video at up to two hundred and forty frames per second. The cleverest part is how it handles remote Linux machines on Wayland, bypassing those annoying permission popups entirely using smart system-level execution. With built-in encrypted peer-to-peer tunneling, you can manage any machine globally without touching port forwarding. Download the client to try lag-free remote control today.

📰 https://news.ycombinator.com/item?id=48993741

🆔 @hackernewsgithubprojects

📦 simoarpe/ziggity

Ziggity

Ziggity is the terminal interface for Git that finally keeps your workspace fast, responsive, and completely keyboard-driven. Built from scratch using Zig, it runs entirely on native Git subprocesses to deliver instant startup times and incredibly low memory use. Instead of taking over your terminal or blocking your screen during background tasks like fetching and pushing, this tool runs operations quietly in the background so you can keep navigating without pause. It even introduces native text selection to copy clean diffs straight from the terminal window. It is the ultimate lightweight upgrade for terminal-focused developers who want absolute control over their Git flow.

📰 https://news.ycombinator.com/item?id=48980506

🆔 @hackernewsgithubprojects

📦 penecho/penecho

PenEcho Infinite AI Canvas

Work through your toughest problems on a massive, zoomable digital whiteboard that actually thinks alongside you. Instead of trying to squeeze your messy thoughts into a tiny chat box, you can freely sketch diagrams, write out math equations, and scribble down handwriting with a stylus or mouse. The system instantly reads your visual layout and drops helpful answers, plots, and even interactive, animated demonstrations directly onto the canvas right next to your work. You can easily lasso your sketches to resize or adjust them before letting the AI see them, keeping your creative flow completely uninterrupted. It is the ultimate spatial scratchpad for visual thinkers.

🆔 @hackernewsgithubprojects

📦 evo-design/proto-tools

proto-tools

Run dozens of complex biological AI models and sequence analysis tools with a single Python installation. Instead of drowning in broken dependencies, conflicting package versions, and GPU configuration headaches, this library manages everything for you. It automatically wraps every single tool—from structure predictors like AlphaFold to advanced protein design pipelines—inside its own isolated virtual environment. You get a clean, unified Python interface to load models, design custom proteins, and analyze genomic data without wrestling with setup scripts. If you are building with biology, search for proto-tools on GitHub and start creating.

📰 https://news.ycombinator.com/item?id=48988050

🆔 @hackernewsgithubprojects

📦 observal/observal

Observal

Stop losing track of your custom AI tools and agents across different coding setups. You can use Observal to centralize all your team's skills, prompts, and local AI servers into a single private registry. Instead of manually writing custom configurations for every developer, this local platform automatically formats your AI components to work with popular tools like Claude Code, Cursor, and Copilot. It also lets you replay entire AI sessions step-by-step, helping you pinpoint exactly where an agent hallucinated or failed. Spin up your own secure registry today and finally see how your team actually uses AI.

📰 https://news.ycombinator.com/item?id=48994984

🆔 @hackernewsgithubprojects

📦 jason8745/llm-agent-trader

LLM Agent Trader

Simulate stock trading with a smart AI assistant looking over your shoulder to explain every single move. The llm-agent-trader repository sets up a complete backtesting dashboard where GPT-4 or Gemini acts as your personal market strategist. Instead of just staring at dry charts, this tool feeds real-time Yahoo Finance data and classic technical indicators directly into a large language model. The AI then spits out clear buy, sell, or hold decisions alongside its step-by-step reasoning and confidence scores. It is the ultimate sandbox for testing how modern AI handles the chaotic swings of the stock market.

🆔 @hackernewsgithubprojects

📦 astaxie/tokenhub

TokenHub

TokenHub is the private enterprise AI gateway that finally puts smart governance and cost control into a single private deployment. By acting as a central bridge between your team and upstream models like OpenAI, Anthropic, or local tools like Ollama, it helps you manage keys, set quotas, and track exact costs by department. Unlike complex cloud gateways, it uses a simple SQLite setup that you can spin up locally in minutes. If you want to give your team safe access to different AI providers without losing track of your budget, this is the perfect self-hosted solution to try today.

📰 https://news.ycombinator.com/item?id=48995071

🆔 @hackernewsgithubprojects