📦 tickernelz/opencode-mem

Give Your AI Coding Agent Persistent Memory

Ever wish your AI coding assistant could actually remember your past projects? This repository introduces a powerful plugin that gives your coding agents persistent memory using a local vector database. By integrating SQLite and high-speed vector indexing, it allows your AI to retain context across different sessions, automatically learn user profiles, and store project memories. It handles complex tasks like smart deduplication and provides a dedicated web interface for you to browse and manage your saved memory timeline. Start building smarter, more context-aware applications today by giving your agents the long-term memory they really need.

🆔 @hackernewsgithubprojects

📦 aws/agent-toolkit-for-aws

Supercharge Your AI Agents with AWS Toolkit

Are you struggling to get your AI agents working effectively with AWS? This official toolkit provides a comprehensive collection of Model Context Protocol servers, specialized skills, and plugins designed to streamline how AI agents build, deploy, and manage cloud infrastructure. By offering structured guides for tasks like optimizing serverless architectures, managing cost billing, or configuring data analytics pipelines, this repository simplifies complex cloud workflows. It is a powerful resource for developers looking to integrate intelligence into their AWS projects seamlessly. Start using these professional tools today to make your AI agents more capable, reliable, and efficient.

📰 https://news.ycombinator.com/item?id=48448344

🆔 @hackernewsgithubprojects

📦 robot-learning-co/trlc-dk1

Build Your Own AI-Native Robot with TRLC DK1

Are you ready to dive into AI-native robotics? The TRLC Developer Kit 1 is an open-source hardware and software platform designed for building and training custom robotic arms. It simplifies the complex world of teleoperation by integrating seamlessly with the LeRobot ecosystem, allowing you to easily record datasets for machine learning tasks. Whether you are performing single-arm teleoperation or complex bimanual recording, this kit provides the CAD files, URDF models, and motor control libraries you need to get moving fast. It is the perfect starting point to bring your own intelligent robot projects to life today.

🆔 @hackernewsgithubprojects

📦 kocoro-lab/kocoro

Turn Your Mac Into an AI Powerhouse

Imagine having an AI partner that lives directly on your Mac, ready to handle your tasks with real computer access. This repository provides a powerful engine and daemon that lets AI agents control your files, apps, browser, and terminal. It solves the friction of switching between tools by allowing you to automate multi-step workflows like researching, coding, or managing documents in one place. With native support for memory and scheduling, it makes complex automation feel natural. Whether you are running CLI commands or building specialized skills, this project bridges the gap between human intent and machine execution.

🆔 @hackernewsgithubprojects

📦 yukiyokotani/office-open-xml-viewer

Build a Custom Office Document Viewer in Your Web App

Ever wanted to render Word documents, Excel sheets, or PowerPoint presentations directly in your own browser-based applications? This unique repository provides a powerful, modular toolkit for exactly that. Built entirely using Rust parsers compiled to WebAssembly, it allows for high-performance rendering onto an HTML Canvas element. The best part is its flexibility: you get headless engines for full UI control, letting you build your own custom viewing experiences rather than relying on standard embeds. Whether you need complex chart support or accurate text layouts, this project offers a professional, self-contained way to integrate Office document viewing seamlessly.

📰 https://news.ycombinator.com/item?id=48436863

🆔 @hackernewsgithubprojects

📦 fsoft-ai4code/codewiki

Automate Your Code Documentation with CodeWiki

Struggling to keep your massive codebase documentation up to date? Meet CodeWiki, an innovative open-source framework designed to generate holistic, architecture-aware documentation for complex, multilingual projects. Instead of basic comments, it uses a recursive multi-agent system to analyze cross-module interactions, producing detailed text and visual artifacts like dependency graphs and sequence diagrams. It supports eight programming languages and handles large repositories by dynamically decomposing them into coherent modules. Whether you use API-based providers or local subscription tools, CodeWiki simplifies repository understanding. Give your project the clear, structured documentation it deserves and start building better today.

🆔 @hackernewsgithubprojects

📦 evolvinglmms-lab/neo

Meet NEO: The Future of Native Vision-Language Models

Ever wondered how we can build truly native vision-language models from the ground up? Meet the NEO series, a groundbreaking project designed to unify pixel-to-word encoding, alignment, and reasoning into one dense, monolithic architecture. Instead of relying on modular components, NEO focuses on a native VLM primitive that delivers superior efficiency. With just a fraction of the data compared to standard models, it achieves powerful visual perception and reasoning capabilities that rival the industry's best. By making these models more accessible and extensible, NEO is paving a new way forward for scalable, high-performance AI. Dive into this ecosystem and explore!

🆔 @hackernewsgithubprojects

📦 tx-leo/humanego

Teach Robots to Perform Tasks Using Only Human Video

Imagine teaching a robot to perform complex tasks just by showing it a few minutes of video from a human perspective. HumanEgo makes this possible by enabling zero-shot robot learning from egocentric recordings. Instead of needing thousands of hours of robot data, this project bridges the gap between human actions and robotic control. It uses vision models to process human hand movements and objects into standardized tokens, allowing robots to replicate those same motions. Whether you are interested in robot policy training or egocentric data collection, this framework provides the tools to turn simple human experiences into actionable robot commands.

🆔 @hackernewsgithubprojects

📦 imran31415/kube-coder

Build Isolated AI Coding Environments in Kubernetes

Are you tired of configuring local development setups that break across different projects? This repository provides a powerful solution using Helm charts to deploy fully isolated, per-user development workspaces directly inside your Kubernetes cluster. Each workspace comes pre-loaded with essential tools like a browser-based IDE, persistent terminal sessions, and even an in-pod browser. By leveraging AI-powered assistants and secure GitHub authentication, it ensures your environment is both highly productive and safe. Whether you need persistent memory or automated build triggers, this project helps you standardize developer workflows while keeping your infrastructure scalable, repeatable, and completely self-contained.

🆔 @hackernewsgithubprojects

📦 boringcollege/zig-by-example

Master the Zig Programming Language Through Practical Examples

Want to level up your systems programming skills? Zig by Example is an incredible resource that helps you master the Zig programming language through a hands-on approach. Instead of just reading theory, you get a collection of annotated, runnable code chapters covering everything from core basics like variables and loops to advanced features like memory allocation, concurrency, and C interop. It cuts through the complexity of modern systems development by showcasing how to write robust, simple code without hidden overhead. Dive into this repository to see real examples in action and start building your own high-performance projects today.

📰 https://news.ycombinator.com/item?id=48444871

🆔 @hackernewsgithubprojects

📦 seraphimserapis/tool-eval-bench

Is Your AI Actually Good at Using Tools? Test It with This Benchmark!

Ever wonder if your AI agent is truly capable of handling real-world tools or just hallucinating its way through tasks? This repository provides a rigorous, deterministic benchmarking suite designed to test exactly how well your LLMs perform at tool-calling. It runs over eighty unique scenarios covering everything from multi-step reasoning and safety boundaries to complex structured outputs. Compatible with popular serving stacks like vLLM, LiteLLM, and llama.cpp, it even includes throughput measurements and pluggable accuracy tests for math and reasoning. Stop guessing about your agentic performance and start measuring it with clear, actionable trace reports to ensure your AI stays on track.

🆔 @hackernewsgithubprojects

📦 ottorenner/gentle-coding

Stop Bullying Your AI: Why Gentle Coding Matters

Are you tired of your AI model giving you strange responses or failing tasks? You might be accidentally stressing it out. Gentle Coding is an open collaborative meta-analysis exploring how human communication affects AI performance. Instead of using high-stress, authoritarian commands that trigger confusion, this project advocates for a balanced, collaborative mindset to keep models within their optimal reasoning boundaries. By implementing simple strategies like defined winning conditions and safety tokens for when the AI hits a bottleneck, you can reduce errors and improve efficiency. Start treating your AI as a tool for success, not a slave, and watch your interactions transform.

🆔 @hackernewsgithubprojects

📦 argahv/sisyphus-academica

Stop Writing Research Papers Alone: Meet the AI Research Swarm

Are you tired of spending weeks grinding through literature reviews and tedious drafting? Meet a revolutionary open-source research pipeline that functions as a self-coordinating swarm of over twenty specialized agents. This system streamlines your academic workflow by managing everything from literature scouting and novelty generation to citation verification and adversarial reviews. It even includes specialized engines to brainstorm unique hypotheses and humanizers to ensure your writing sounds natural rather than machine-generated. Whether you need help with technical formatting or rigorous peer critique, this tool is designed to save you massive amounts of time. Elevate your research game and start automating today.

🆔 @hackernewsgithubprojects

📦 zenalexa/toki-bitemporal-memory

Stop LLM Memory Contradictions with TOKI

Ever notice how LLM agents get confused when they receive conflicting information? This repository introduces TOKI, a clever bitemporal operator algebra designed to solve contradiction resolution in persistent agent memory. By treating belief updates as versioned writes, it uses four specialized operators to keep your data sound and audit-ready. The project includes a full experimental harness that proves why many production systems fail to prevent common anomalies like replay inconsistency. If you are building reliable LLM applications, check out this reference implementation to see how you can finally keep your agent's memory organized, consistent, and fully auditable.

🆔 @hackernewsgithubprojects

📦 amarnath3003/mcpify

Turn Any App Into an AI Agent with MCPify

Ever wish your existing software could just talk to AI agents? Modern apps are built for humans, but MCPify acts as an AI enablement compiler to change that. It scans your codebase, including backend services, frontend actions, and database schemas, to automatically generate a functional, permission-aware Model Context Protocol server. This allows agents to safely interact with your workflows without brittle browser automation or manual tool building. It even detects multi-step processes and adds a security layer to keep operations safe. Compile your stack today and give your software the power to work natively with autonomous agents.

🆔 @hackernewsgithubprojects

📦 arturr0k3r/akiraos

AkiraOS: The Future of Embedded Development with WebAssembly

Are you tired of reflashing your entire device just to update a small feature? AkiraOS is changing the game for microcontrollers by bringing a container-like experience to your hardware. Built on top of the powerful Zephyr RTOS, it allows you to run sandboxed applications as portable WebAssembly binaries. This means your core firmware stays stable while you update your apps over the air. With features like the Capability Guard for granular security and AOT compilation for near-native performance, AkiraOS makes building robust IoT devices faster and safer. Experience the freedom of modular, secure, and truly portable embedded development.

📰 https://news.ycombinator.com/item?id=48453235

🆔 @hackernewsgithubprojects

📦 deusdata/codebase-memory-mcp

Supercharge Your AI Coding Agents With Codebase Memory

Stop wasting tokens on slow file searches and upgrade your AI coding workflow with Codebase Memory. This high-performance engine indexes your entire repository into a searchable knowledge graph in just milliseconds, supporting over one hundred fifty languages. By replacing repetitive file scans with precise structural queries, it slashes token usage significantly while boosting answer quality for your favorite AI coding agents. It runs locally as a single, dependency-free binary on Windows, macOS, and Linux, providing instant architectural insights. Simplify your development process, reclaim your token budget, and give your AI agents the deep codebase understanding they truly need.

🆔 @hackernewsgithubprojects

📦 vava-nessa/free-coding-models

Find the Fastest Free AI Coding Models Instantly

Are you tired of searching for reliable free coding models? This powerful tool tracks over 150 models across 16 providers in real time, saving you the headache of guessing which one is actually fast. It performs live benchmarks, calculating a stability score based on speed, latency, and uptime, so you only use the best options. Once you pick a model, it automatically configures your favorite coding tools for you, letting you start building in seconds. Stop wasting time hunting for the right AI and start coding faster today with this essential utility for every developer's terminal setup.

🆔 @hackernewsgithubprojects

📦 freestyle-voice/freestyle

Type Faster With Your Voice Using This Open Source App

Tired of slow typing? Imagine speaking your thoughts and seeing them appear instantly as clean, polished text right where your cursor is. Freestyle is an open source, local-first dictation app designed to make your workflow significantly faster. It solves the friction of manual typing by allowing you to dictate through a simple hotkey, while offering advanced features like automatic grammar cleanup and custom vocabulary replacements. Because it is local-first, your private voice data never leaves your device. Whether you need to draft emails or write notes, this powerful tool lets you communicate at the speed of thought.

🆔 @hackernewsgithubprojects

📦 atomicstrata/atomicmemory

Give Your AI Agents a Real Memory Layer

Are you tired of your AI agents forgetting everything between sessions? AtomicMemory solves this by providing a portable, inspectable semantic memory layer you can embed directly into your code. Instead of relying on a black-box service, you get full control over how your agents store, retrieve, and revise facts over time. With a TypeScript SDK, framework adapters for tools like LangChain, and a dedicated CLI, it bridges the gap between local privacy and powerful context management. You own your data, ensuring your agents grow smarter without losing the plot. Give your AI a reliable memory today and start building better agents.

🆔 @hackernewsgithubprojects

📦 lollokara/bugbuster

Give AI Physical Hands with BugBuster

Ever wanted your AI to debug physical hardware? Meet BugBuster, an open-source platform that bridges the gap between digital intelligence and the electronics bench. By using a Model Context Protocol server, it gives AI assistants like Claude direct control over a dedicated hardware board. It handles everything from measuring voltages and capturing waveforms to driving outputs and debugging embedded targets. With sixty-one AI-callable tools, you can finally automate complex diagnostic tasks that once required manual probing. Whether you are analyzing digital signals or managing power rails, BugBuster turns your AI into an expert hardware engineer. Start automating your lab today.

🆔 @hackernewsgithubprojects