Show HN: Vpuna AI Search – A semantic search platform https://ift.tt/Mftw5r6

Show HN: Vpuna AI Search – A semantic search platform Dear HN Community, I am a long time fan and first-time contributor. I just launched a developer focused semantic search platform and wanted to share it with the community. The idea is simple: upload structured or unstructured documents, select the fields you want to index and tag as metadata, and instantly get a clean search API you can use in your own app. Here is what it currently supports: - Manage your own tenants and projects - Upload .json and .txt files (support for .pdf, .docx, .xlsx, .yml, etc. coming soon) - Expose 3 APIs: search, upload document (embeddings), and delete document - Manage your own API keys - Uses CPU based sentence-transformers/all-MiniLM-L6-v2 for embeddings ( support for other local and online models are coming soon) LLM summarization and Model Context Protocol (MCP) support are on the roadmap Why I built it: In my consulting work, I kept seeing client wanting to move beyond basic keyword search and integrate semantic search with optional summarization. Most existing tools are either too expensive, too restrictive, or require custom layers (like custom Python servers for pre processing queries and embeddings). I wanted something API first, developer friendly, and easy to self host or use out of the box. This is the first release, and I would love your feedback. Would you use this? What is missing for your use case? Here is the README with all the links https://ift.tt/FgWhS1O Thank you for your time. https://ift.tt/OLc8kq9 June 20, 2025 at 11:24PM

Show HN: RM2000 Tape Recorder, an audio sampler for macOS https://ift.tt/X2oNpmG

Show HN: RM2000 Tape Recorder, an audio sampler for macOS RM2000 Tape Recorder makes it stupid simple to grab audio samples and organize them: just record the sample, give it a title (and maybe some tags), and it is saved neatly into a directory of your choosing. I'm a huge datahoarder and have always appreciated tools / services like PureRef and Are.na which help me make sense of everything I collect. Those services concern themselves with images and video - I wondered, why can't the same be done with music and audiofiles? I actually got the inspiration for the filenaming scheme from the Emacs Denote package - every sample is saved in the format of title--tag1--tag2.mp3. Emacs Denote does something similar, for example an identifier--title--keywords.org . I chose this method as any file browser with fuzzy search can search through samples, i.e. - the Ableton file browser. Just search up some of the tags, and a title, and you'll be able to find your sample. I wanted this app to look good, as well (and is why I spent so much time making it!) The app is made with a mix of SwiftUI and AppKit, while the assets were rendered in Sketch I appreciate your time and I'd love to hear your thoughts on it. If you do download it, and find suggestions / bugs, please let me know! Cheers https://rm2000.app June 17, 2025 at 09:50PM

Show HN: Relix: A Unix-like OS based on MIT's xv6 https://ift.tt/pRtFIyw

Show HN: Relix: A Unix-like OS based on MIT's xv6 Hello everyone, this is my first post as someone encouraged me to post this here. I have been working on Relix for over a year and am willing to answer any questions you may have! https://ift.tt/aql9pUR June 20, 2025 at 12:53AM

Show HN: AI Debate Arena – See Which LLM Argues Best https://ift.tt/94OmjDu

Show HN: AI Debate Arena – See Which LLM Argues Best Ever wish you could get the best arguments for both sides of a debate? I built an AI-powered debate platform that pits language models against each other on controversial topics. Each AI is randomly assigned a side (pro/con). You vote before and after to see if you were persuaded. Most content today presents lopsided arguments. They provide strong points for one side, weak ones for the other. This project aims to surface the strongest arguments from both sides, using LLMs to simulate a fair debate. With enough usage, I want to use it to benchmark LLMs. My hypothesis is that randomly assigning sides of the debate, models with built-in biases will score worse. It’s currently using GPT 4o, Grok 3, and Gemini 2.5 Flash. It’s early, still rough around the edges, and I’d love feedback on the concept and direction. Curious how the HN crowd thinks this could evolve. It’s built for the intellectually curious that are open minded about changing their positions. Some next steps I’m considering: - Tuning the length and structure of arguments - Prompting improvements to reduce rhetorical fluff - Optional audio output of debates Try it out and let me know what you think! https://ift.tt/wU6QRA2 June 19, 2025 at 01:56AM

Show HN: Turn long form videos into short form clips https://ift.tt/PyKjgk8

Show HN: Turn long form videos into short form clips https://ift.tt/iczmNZt June 18, 2025 at 11:22PM

Show HN: I couldn't poop, so I built an app to track digestion in real-time https://ift.tt/bgUfWBq

Show HN: I couldn't poop, so I built an app to track digestion in real-time https://ift.tt/ZNs2qSM June 19, 2025 at 12:02AM

Show HN: Superscan – Visualize filetree for filesystem, gdrive, S3 buckets etc. https://ift.tt/vKgmadI

Show HN: Superscan – Visualize filetree for filesystem, gdrive, S3 buckets etc. https://ift.tt/U9H8IQt June 18, 2025 at 02:52AM

Show HN: PMDb – Public Movie Database https://ift.tt/yWIOqoA

Show HN: PMDb – Public Movie Database https://pmdb.dev/ June 18, 2025 at 12:07AM

Show HN: I built a FOSS tool to run your Steam games in the Cloud https://ift.tt/rT6pEGs

Show HN: I built a FOSS tool to run your Steam games in the Cloud I wanted to play my Steam games but my aging PC couldn’t keep up, so I built Cloudy Pad - a tool to run Steam in the Cloud (GitHub: https://ift.tt/nQlY5EZ ) It runs on AWS, Azure, GCP, Scaleway and Paperspace with various cost optimizations and safeties: - Cost alerts - Auto stop inactive instances to avoid unwanted cost - Disk snapshots and data cleanup for cost efficiency - Spot instance support Under the hood: a Linux VM and a container running Sunshine (a streaming server https://ift.tt/NAIWuXo ) with Steam. Most Windows games work just fine thanks to Proton. It streams effortlessly at 1080p 100+ FPS - I recently played Baldur’s Gate III and Clair Obscur in Ultra, ran like a breeze. Cost-wise it’s great for occasional players: ~30h or less per month typically cost less than 25$. Though admittedly for heavy gamers it may be less cost-effective due to cloud pricing. I’d love feedback from the HN community ! https://ift.tt/nQlY5EZ June 18, 2025 at 12:27AM

Show HN: Nexus.js - Fabric.js for 3D https://ift.tt/6bU0kFr

Show HN: Nexus.js - Fabric.js for 3D I was looking for a tiny library to easily transform both 2D & 3D objects with simple mouse / touch controls and a fixed camera, in the browser. Like a simple 3D editor but without requiring the user to be a Blender expert. Couldn't find anything lightweight, so I’m building one. Think Fabric.js but for 3D. Built entirely with Three.js / R3F. Borrowed some inspiration from VR/AR interaction systems for controls. Feel free to play with it and let me know what you think! https://ift.tt/RjUCtZv June 17, 2025 at 02:03AM

Show HN: Drop – Svelte like compiler for Web components https://ift.tt/Oj0PQb3

Show HN: Drop – Svelte like compiler for Web components I've built a small compiler, heavily inspired by Svelte, that leans on modern web standards and proposals, namely Web Components, HTML Modules, and Signals. Although web components never really took off, I still believe they have strong potential as a foundation for building web applications without relying on a framework. GitHub: https://ift.tt/vjJPoQz Blog post: https://ift.tt/iJn4I1b... I’d appreciate some feedback before committing more time to this project ! https://ift.tt/vjJPoQz June 17, 2025 at 12:25AM

Show HN: Canine – A Heroku alternative built on Kubernetes https://ift.tt/nEe6PSL

Show HN: Canine – A Heroku alternative built on Kubernetes Hello HN! I've been working on Canine for about a year now. It started when I was sick of paying the overhead of using stuff like Heroku, Render, Fly, etc to host some web apps that I've built. At one point I was paying over $400 a month for hosting these in the cloud. Last year I moved all my stuff to Hetzner. For a 4GB machine, the cost of various providers: Heroku = $260 Fly.io = $65 Render = $85 Hetzner = $4 (This problem gets a lot worse when you need > 4GB) The only downside of using Hetzner is that there isn’t a super straightforward way to do stuff like: - DNS management / SSL certificate management - Team management - Github integration But I figured it should be easy to quickly build something like Heroku for my Hetzner instance. Turns out it was a bit harder than expected, but after a year, I’ve made some good progress The best part of Canine, is that it also makes it trivial to host any helm chart, which is available for basically any open source project, so everything from databases (e.g. Postgres, Redis), to random stuff like torrent tracking servers, VPN’s endpoints, etc. Open source: https://ift.tt/Ir2jHpw Cloud hosted version is: https://canine.sh https://ift.tt/Ir2jHpw June 16, 2025 at 11:57PM

Show HN: Pipo360 – Generate production-ready back end APIs in 60 seconds with AI https://ift.tt/5l2EDWg

Show HN: Pipo360 – Generate production-ready back end APIs in 60 seconds with AI Hi HN , I got tired of writing the same boilerplate over and over — DB setup, auth, routes, security — every time I built a backend. So I built Pipo360 — an AI-powered tool that generates production-ready backends in under 60 seconds, from just a plain-text description. How it works: Type what you need “Create a task management API with user auth and MongoDB” Hit Generate Get real, exportable code Auth (JWT) Database schema CRUD routes Deployable to Vercel, AWS, etc. No templates. No lock-in. Just code that works. Why it’s different: Built with Gemini AI + human supervision (to ensure real prod-quality output) Exports to MongoDB, PostgreSQL, MySQL, SQLite Secure by default (JWT, RBAC, etc.) Supports no-login backend previews Try it live (No signup needed): https://pipo360.xyz Would love feedback: What backend would you try first? What would make it better for your workflow? Would open sourcing part of it be useful? https://pipo360.xyz June 16, 2025 at 01:16AM

Show HN: Seastar – Build and dependency manager for C/C++ with Cargo's features https://ift.tt/Fb0YQdp

Show HN: Seastar – Build and dependency manager for C/C++ with Cargo's features Hi hackers! I'm a self-taught solo teenage dev working on Seastar, a unified build system and dependency manager for C and C++. It is capable of compiling and linking projects, managing recursive dependencies and headers, and even has a template system -- your C++ library is one `seastar new mylib --lang c++ --lib` away! Also, everything is configured in TOML, because TOML is awesome. C is one of my favorite languages, but I usually end up writing stuff in Rust because I love Cargo. Unlike C, Cargo handles the dependencies, linking, globbing, and so much more for you. So I wrote Seastar to give that function in C and C++. What's planned? A package registry like crates.io, compatibility with CMake projects, commands to migrate, and so much more. If you have more ideas, please give them! I am trying to reach 150 stars by the end of summer, and thus a star would be greatly appreciated! This project is still in development, and a star helps out a ton. https://ift.tt/MiBONcf June 16, 2025 at 01:06AM

Show HN: Get nutrition labels for any recipe (YouTube, Website, Text, Image) https://ift.tt/yUFwXbv

Show HN: Get nutrition labels for any recipe (YouTube, Website, Text, Image) Heya HN, I love watching cooking videos on YouTube, and one day an idea sparked: Can I instantly get the nutritional values for this recipe? The problem: Great recipes are everywhere, but figuring out the actual nutrition is a chore. Most of us who track calories or macros have to: * Manually get nutritional info for every ingredient. * Wrestle with spreadsheets for calculation. * Or just give up and eat that lasagna. Let’s be real: no one enjoys that, especially when you just want to cook and eat your food. So I built Recp.ai. It’s a free tool that pulls the nutrition data for you. Just give it a YouTube link, a recipe from a website, a photo of a cookbook page, or even just pasted text. It identifies the ingredients and quantities, matches them against the USDA database, and gives you a full nutrition label. It started as a script to pull ingredients from YouTube transcripts using Gemini. Then I got obsessed. Why not any website? So I added a scraper. What about any list of ingredients? Added text parsing. How about a cookbook? Now it uses Google Cloud Vision so you can just upload or snap a photo of the recipe. I wanted to build something lightweight, fast, and simple that you'd actually use. No sign-ups. No ads. Privacy-first. What I’m happy with? It works on a huge variety of sources. I fed it a photo of my old Escoffier cookbook recipe and voila - it works. What's next? I'm planning to use Google Cloud Vision to identify the dish, say "Beef Pho", it'd figure out its typical ingredients, and generate an estimated nutrition label. It would then ask users to confirm the dish, so the result is as accurate as possible. Any suggestions here if this would be the right way? Would love to hear your thoughts, and if this feels like something that would with your meal prep. Just sharing something I built because I wanted it to exist and solve a problem. June 16, 2025 at 12:23AM

Show HN: Tikt.com – Remove the "OK" from TikTok URL's to Download as MP3 or MP4 https://ift.tt/kHFRY7o

Show HN: Tikt.com – Remove the "OK" from TikTok URL's to Download as MP3 or MP4 https://tikt.com/ June 15, 2025 at 11:32PM

Show HN: ZeroConfigDNLA – Easy to run media server in Python https://ift.tt/VHiz8py

Show HN: ZeroConfigDNLA – Easy to run media server in Python The goal was to be able to serve videos from my laptop in one command. Give it a go and let me know if it works for you! If you run into issues, please provide log output and the source and destination device info (make/model/etc) https://ift.tt/FrZGIB8 June 15, 2025 at 03:46AM

Show HN: S3mini(v0.2) – Basic S3 Support for Ceph and Oracle Object Storage https://ift.tt/NdK0RFP

Show HN: S3mini(v0.2) – Basic S3 Support for Ceph and Oracle Object Storage https://ift.tt/M10aroS June 15, 2025 at 04:18AM

Show HN: AnyCrawl v0.0.1-alpha.5 – custom user-agent and richer scraping API https://ift.tt/OxJ0clw

Show HN: AnyCrawl v0.0.1-alpha.5 – custom user-agent and richer scraping API ## [0.0.1-alpha.5] - 2025-06-14 ### Added - Integrated AWS S3 storage support with new `S3` class and environment variables for seamless file uploads and retrievals. - Introduced `FileController` for serving files from S3 or local storage with robust path validation and error handling. - Added multiple content transformers (Screenshot, `HTMLTransformer`) improving HTML/Markdown extraction and screenshot generation. - Extended scraping capabilities with new options: output `formats`, `timeout`, tag filtering, `wait_for`, retry strategy, viewport configuration, and custom user-agent support. - Added Safe Search parameter to `SearchSchema` for filtered search results. - Refactored engine architecture with a factory pattern and new core modules for configuration validation, data extraction, and job management. - Implemented graceful shutdown handling for the API server and improved logging for uncaught exceptions / unhandled rejections. - Added Jest configuration for API and library packages with ESM support and updated test scripts. - Updated CI workflows to publish Docker images on version tags. - Expanded README with detailed environment variable descriptions and API usage examples. ### Changed - Refined error handling in `ScrapeController` and `JobManager`; failure responses now include structured error objects and HTTP status codes. - Enhanced `BaseEngine` with explicit HTTP error checks and resilience improvements. - Updated OpenAPI documentation to reflect new scraping parameters and error formats. - Migrated key-value store name to environment configuration for greater flexibility. - Enhanced per-request credit tracking in `ScrapeController` and enhanced logging middleware to include credit usage. ### Fixed - Improved job failure messages to include detailed error data, ensuring clearer debugging information. - Minor documentation corrections and clarifications. https://ift.tt/czTLNun June 14, 2025 at 11:18PM

Show HN: I built a Mac app to restore Dock-click minimize and avoid tiny buttons https://ift.tt/YaRM7Ud

Show HN: I built a Mac app to restore Dock-click minimize and avoid tiny buttons Hey HN, I'm the developer behind Click2Minimize. This app is my personal fix for two long-standing frustrations with the macOS interface. First, I wanted to restore Dock-click minimize. On other operating systems, I was used to clicking an app's icon to minimize its window—a simple, fast toggle. On a Mac, that second click does nothing, which always felt like a dead end in my workflow. Second, I was tired of having to deal with the tiny buttons. So much of window management—minimizing, maximizing, arranging—forces you to stop what you're doing, carefully aim your cursor at one of three small dots, and click. It's a constant micro-interruption. The Solution: A Fluid, Mouse-First Approach ----------------------------------------------------- Click2Minimize is a lightweight, native utility that turns your entire window title bar into a powerful gesture area. The goal is to let you manage your workspace without ever needing to aim for those little dots. * Consistent Dock Behavior: Click on Dock icon to minimize/hide the app. * Minimize Window Under Mouse: Simply hold down left mouse button and click the right one, or double-click the right button. * Maximize Window Under Mouse: Simply hold down right mouse button and click the left one, or double-click the notch area. * Snap Window to Left/Right: Simply hold down right button and rock the scroll wheel, or use fn key while swipe on trackpad. * Restore Window Size & Position: Holde down right button and click middle button, or user fn key with right-click on trackpad. * And many other useful gestures, such as the App Switcher and changing workspaces, were also included. Most importantly, it handled macOS full-screen mode smoothly and no longer felt intrusive. It is designed to resemble a missing feature of the operating system, with all gestures being highly intuitive, especially when using a mouse, as there is no need to remember keyboard shortcuts or bring the window to the front. Feedback, Discount & Free Licenses: ---------------------------------------- I'm here all day and would love to hear your thoughts. I genuinely want to make this app better, and the HN community's feedback is invaluable. Furthermore, I'll be sending a completely free license to the commenters with the most thoughtful feedback, bug reports, or feature suggestions I see. You're not just buying an app; you're helping to shape it. Link: https://ift.tt/JjowNtS https://ift.tt/JjowNtS June 14, 2025 at 11:51PM