📦 basketikun/infinite-canvas

infinite-canvas

Design and iterate on your visual ideas in a single place by organizing a visual workspace that connects to any open A-I compatible model. This open-source platform gives you an infinite canvas to arrange image generation, picture editing, and video creation side-by-side. Instead of jumping between tabs, you can drag nodes, connect ideas, and run text-to-speech or generation scripts straight from your browser. The coolest part is how it lets you run a local helper agent to control the canvas using natural language commands. It is a fantastic playground for developers wanting to map out creative workflows without getting bogged down in complex setups.

🆔 @hackernewsgithubprojects

📦 lightseekorg/tokenspeed

TokenSpeed

TokenSpeed is the high performance inference engine that finally gives you blazing fast speed for agentic workloads without the headache. It acts as a supercharged control center for running large language models, matching the top tier performance of industrial tools while remaining incredibly easy to use. The project uses a clever compile time scheduler to safely manage memory, and it packs some of the fastest custom attention kernels in the world. It already powers massive models with record breaking speeds, making it perfect for anyone deploying complex AI agents. Check out the project to accelerate your local AI setups today.

🆔 @hackernewsgithubprojects

📦 alibaba/skill-up

skill-up

Evaluate and evolve your AI agent's capabilities through natural conversation instead of wrestling with messy, ad-hoc test folders. This clever tool lets you define test cases in clean, declarative YAML files and run them across multiple agent engines, including Claude Code, Codex, and custom setups. When a test fails, its companion agent, skill-upper, reads the report, automatically diagnoses the failure, repairs or expands your test suite, and runs the evaluation loop again. You get structured, continuous-integration-ready reports detailing exactly where your agent is falling short and how it is improving. It is like having a dedicated QA engineer constantly hardening your AI's custom skills.

🆔 @hackernewsgithubprojects

📦 lumenastrum/clio-style-preview

Clio Style Preview

Instantly preview hundreds of artistic styles directly inside ComfyUI before generating a single image. Clio Style Preview gives you a library of nearly four hundred dense prose prompt styles, complete with a visual style picker and search engine built right into your node graph. Instead of guessing what a prompt like Yoshitaka Amano looks like, you get live, gorgeous thumbnails on one subject with a single seed. You can even batch render your own subject through the entire library and use a brilliant split slider gallery to watch one artistic style melt into another. It is the perfect playground for fine tuning your AI art generations.

🆔 @hackernewsgithubprojects

📦 do-md/domd

Meet domd, the 20KB Markdown Engine built for AI

domd is the Markdown editor that processes twenty-thousand-line documents with absolutely zero lag. Instead of piling on heavy rich-text frameworks, this project uses a custom, twenty-kilobyte native engine where the Markdown file itself remains the single source of truth. The coolest part is how it handles streaming AI output. When an AI model spits out token-by-token text, domd ingests those chunks on the fly, rendering half-built tables and open code blocks instantly without any distracting screen flicker. It even handles fine-grained, offline collaborative edits within a single paragraph. It is fast, lightweight, and perfect for the AI era.

🆔 @hackernewsgithubprojects

📦 agentrhq/webcmd

Webcmd

AI agents can cut their browser token spending by up to ninety percent using an intelligent self-learning infrastructure. Instead of forcing your artificial intelligence to waste computing power rediscovering the exact same websites and layouts on every single run, webcmd remembers how they work. The system learns the structure, actions, and navigation paths of the sites your agents visit, converting those observations into highly efficient, reusable commands. It acts like a custom shortcut system for the web, giving agents direct, deterministic access to things like search results or social media bookmarks. The next time your agent needs to complete a task, it bypasses heavy manual browsing altogether, saving you massive...

🆔 @hackernewsgithubprojects

📦 cdinnison/ray-finance

Ray Finance

Turn your raw bank data into actionable financial moves. Ray finance connects to your real accounts, analyzes your unique family situation, and guides you on exactly what to do next rather than just showing you generic pie charts. It operates locally with an encrypted database to keep your personal details completely private, and it actually remembers your goals and updates your profile automatically over time. You can even run a fully realistic demo mode right from your command line to test out the smart financial advisor before linking your own bank. Give it a spin to see how clean and intelligent local budgeting can be.

🆔 @hackernewsgithubprojects

📦 ctrlb-hq/ctrlb-decompose

ctrlb-decompose

An engineering team can compress over one million messy server log lines down to just a few dozen clear, structured patterns in a single pass. The tool ctrlb-decompose does exactly that by analyzing raw text stream inputs, grouping repeating lines, and extracting variables like IP addresses or durations on the fly. It automatically calculates percentages and points out anomalies, such as weird timing spikes, without heavy database setups. You can output these structured patterns into terminal-friendly charts, clean JSON, or highly compressed markdown files tailored for large language models. Try running it locally to clean up your log files before sending them to your AI assistant.

📰 https://news.ycombinator.com/item?id=49083336

🆔 @hackernewsgithubprojects

📦 scaccogatto/okf-skills

okf-skills

Your AI coding agents can now automatically build and verify an interactive, visual map of your team's collective knowledge directly inside your codebase. The project okf-skills brings the Open Knowledge Format to Claude Code and other agents, letting them turn your scattered markdown files into a structured, queryable knowledge graph. Instead of wasting time on missing context, this tool uses a deterministic validator to ensure your documentation remains accurate and up to date, while compiling everything into a self-contained visual map that lets you explore how your projects, runbooks, and decisions link together. It is the easiest way to give your development team and your AI assistants a single, shared...

🆔 @hackernewsgithubprojects

📦 saad-khan-rind/nosatmosphereeffect

nosatmosphereeffect

Transform your boring Android home screen into a dynamic work of art using gorgeous custom transition effects. If you love the signature lock screen transitions from Nothing operating system but do not own one of their phones, this clever app lets you port that exact magic to any Android thirteen plus device. It smoothly shifts your wallpaper from a moody, blurred abstract cloud view into a crystal clear image the second you unlock your phone. You can even try alternative styles like frosted glass, liquid color flows, or retro comic book dots. Give your phone a premium, interactive feel today.

🆔 @hackernewsgithubprojects

📦 jaypokale/chisle

Chisle

Chisle can cut your artificial intelligence coding assistant's output token bill nearly in half with a single install command. While most prompt compressors only trim conversational fluff, this tool targets output prose, boilerplate code, and oversized input context all at once. It injects a clever ruleset that teaches models to reuse existing code, leverage standard libraries, and write the absolute minimum needed to solve a problem. It even uses a built-in hook to scrub, elide, and deduplicate noisy command outputs before they can bloat your session context. It is the ultimate downside insurance to keep your development runs fast, focused, and incredibly cheap.

🆔 @hackernewsgithubprojects

📦 verilean/sparkle

Sparkle HDL

Sparkle is a compiler that lets you write computer chips and hardware designs using the Lean programming language, prove they are completely free of bugs, and then turn them into standard Verilog code. Normally, writing hardware is a stressful process of guessing and testing to avoid physical and logical errors. Sparkle changes this by letting you define exactly how your hardware should behave as pure mathematical functions, prove its safety using Lean's built-in theorem prover, and then generate clean, readable SystemVerilog. It represents a massive leap forward for designing incredibly secure, high-assurance computer chips.

📰 https://news.ycombinator.com/item?id=47569547

🆔 @hackernewsgithubprojects

📦 jgbrwn/vibebin

Vibebin

Spin up secure, personal coding sandboxes directly on your own virtual private server to run artificial intelligence coding agents in complete isolation. This project coordinates lightweight Linux containers on your hardware, pre-configuring them with developers' favorite tools and multiple automated coding assistants. Instead of relying on expensive cloud subscriptions, you get a local dashboard to launch, pause, and snapshot your environments instantly. It handles the networking, reverse proxying, and secure shell routing automatically, letting you connect your desktop editor or access a web interface with a single command. It is the ultimate self-hosted playground for safe, isolated experimentation.

📰 https://news.ycombinator.com/item?id=47366050

🆔 @hackernewsgithubprojects

📦 thenetworkchuck/dark-web-scraping-guide

Dark Web Scraping Guide

About ninety percent of the dark web is made of fake sites, scams, and law enforcement honeypots, meaning security analysts historically had to spend up to eight hours wading through junk just to find real threat intelligence. The dark web scraping guide repository changes the game by introducing Robin, an artificial intelligence powered research tool that slashes that manual search time down to just thirty minutes. By querying fifteen dark web engines at once, it uses semantic artificial intelligence to filter hundreds of sketchy results into a clean list of verified sources, scrapes them in parallel, and generates clean markdown reports. It is the ultimate way to safely study real-world...

🆔 @hackernewsgithubprojects

📦 bitwize-ai/logue

Logue

Your Mac can now transcribe meetings and edit documents with total privacy, running advanced artificial intelligence completely on device. Logue is an open-source assistant built for Apple Silicon that handles real-time audio transcription, speaker labeling, and structured meeting summaries without ever sending your data to the cloud. By running local models directly on your hardware, it keeps your conversations, action items, and notes entirely offline and secure. It also features a rich text editor that provides inline grammar corrections, tone adjustments, and factual verification. It is the ultimate tool for professionals who need powerful writing and meeting assistance without sacrificing their privacy.

📰 https://news.ycombinator.com/item?id=49007801

🆔 @hackernewsgithubprojects

📦 miaai-lab/glm-5.2-nvfp4-aqlm-triple-dgx-sparks

glm-5.2-nvfp4-aqlm-triple-dgx-sparks

The glm-5.2-nvfp4-aqlm-triple-dgx-sparks deployment recipe is the local serving stack that finally lets you run a massive seven hundred and fifty-three billion parameter model right from a desktop cluster. Instead of relying on massive cloud data centers, this clever setup combines highly compressed weight formats to pack the entire model onto just three compact machines. It solves the headache of extreme hardware demands by squeezing a frontier-grade model into a fraction of its original memory footprint, serving up a huge context window with impressive generation speeds. If you want to break free from cloud subscription monopolies and run giant artificial intelligence locally, this setup is your ultimate roadmap.

🆔 @hackernewsgithubprojects

📦 krishagarwal314/autodev-studio

autodev-studio

Describe a new software feature in plain English and watch a team of AI agents build, test, and ship it directly to your repository. Instead of sending an expensive AI model blindly into a massive codebase to hunt around and waste tokens, autodev-studio builds a localized knowledge base of your code just once. It uses static analysis to create a map of your system and coordinates different specialized AI agents for planning, writing code, and reviewing. You can mix and match models, allowing one model family to write the code while an entirely different one reviews it to eliminate bias. It is a brilliant way to ship verified pull requests...

🆔 @hackernewsgithubprojects

📦 letsseal/letsseal

Let's Seal

Let's Seal is the document signing engine that makes proving any file is real completely free and open source. Think of it as the Let's Encrypt of document proof, ending the era of expensive, rented digital seals. When you seal a file, it cryptographically guarantees the document is unaltered, timestamps it to the Bitcoin ledger, and binds it to a verified domain. Whether you are dealing with PDF contracts, images, code, or emails, the seal rides along inside the file in its native format. This means anyone can verify your documents offline using standard tools without ever needing an account or trusting a middleman.

📰 https://news.ycombinator.com/item?id=49071365

🆔 @hackernewsgithubprojects

📦 luce-org/lucebox

Lucebox

Lucebox is the local AI server that finally lets you run massive language models at blisteringly fast speeds on standard consumer graphics cards. If you want to run high-end open models on a single graphics card like an older RTX 3090, you are usually stuck waiting for slow, painful generation speeds. This project completely solves that bottleneck by combining custom GPU-accelerated code, smart speculative decoding, and clever context compression. It acts like a high-octane engine for your local AI chat clients, squeezing up to eight times more performance out of your existing hardware. It is the ultimate tool for running supercharged, private AI right at home.

🆔 @hackernewsgithubprojects

📦 cynative/cynative

Cynative

Audit your entire cloud infrastructure using artificial intelligence without risking any accidental modifications. Cynative connects frontier language models directly to your code, cloud, and active runtime systems to reason through everything as a single unified environment. It solves the massive headache of tracking down security risks by automatically writing and executing read-only scripts in a secure sandbox. This lets you ask simple questions in plain English, and the system parallel-queries your cloud providers to trace vulnerabilities back to the exact code changes that caused them. Give it a try to find hidden security gaps across your systems instantly.

📰 https://news.ycombinator.com/item?id=49086558

🆔 @hackernewsgithubprojects