Show HN: Routing24 – Free route optimization in the browser Hi HN! I built Routing24 to make route optimization easier and accessible for small businesses and solo drivers. With Google Maps, you can plan routes between a few locations for free, but it doesn’t support efficient multi-stop planning, handling multiple vehicles, or adding specific business rules like delivery time windows. Many tools offer route optimization for around $30 per vehicle per month, but Routing24 provides it completely free. Optimization happens fully on the client-side, using your device’s resources instead of cloud servers. The interface is simple for now: you can import/export using Excel, and there’s a web UI to edit and view imported data and optimized routes. Data is saved or uploaded anywhere, and one browser tab runs one optimization. Map data comes from OpenStreetMap, and the solver is built in C++/WASM, using public domain algorithms. I’d love to hear any feedback as I keep developing it. Thanks, and I hope it’s useful! https://routing24.com October 30, 2024 at 08:06PM
Show HN: Modus, serverless framework for intelligent APIs powered by WebAssembly Modus is an open-source, serverless framework for building APIs powered by WebAssembly. It simplifies integrating AI models, data, and business logic with sandboxed execution. We built Modus to put code back at the heart of development. You write a function. export function sayHello(name: string): string { return `Hello, ${name}!`; } Then, Modus: - extracts the metadata of your functions - compiles your code with optimizations based on the host environment - caches the compiled module in memory for fast retrieval - prepares an invocation plan for each function - extracts connections, models, and other configuration details from the app’s manifest - generates an API schema and activates the endpoint You query the endpoint. query SayHello { sayHello(name: "world") } In a few milliseconds, Modus: - loads your compiled code into a sandboxed execution environment with a dedicated memory space - runs your code, aided by host functions that power the Modus APIs - securely queries data and AI models as needed, without exposing credentials to your code - responds via the API result and releases the execution environment Now you have a production ready scalable endpoint for your AI-enabled app. AI-ready when you’re ready. Launch and iterate. Read more in the announcement blog post: https://ift.tt/e2g5A1w... Come join our Discord, we’d love to hear your feedback: https://ift.tt/R0LrbUH Follow us on GitHub: https://ift.tt/il2Q4GT Modus docs: https://ift.tt/MhgN4zf Modus quickstart video: https://youtu.be/3CcJTXTmz88 https://ift.tt/il2Q4GT October 30, 2024 at 09:32PM
Show HN: Let's Play (Voice AI) Ball – A Baseball Strategy Game My 8 year old son is obsessed with baseball. It’s unfortunate because 1) I deeply love lacrosse (I was the leading scorer my sophomore year for UC Davis! Never miss an opportunity to say it!), 2) I don’t know much about baseball, and 3) he’s a diehard Yankees fan. Although I can’t do anything about his team affiliation or the current 3-0 Yankees deficit in the 2024 World Series, I can do something about teaching him baseball strategy -- despite knowing little about baseball myself. I built this Voice AI game for him to learn more about baseball strategy. In the game, the Voice AI describes a baseball scenario, and the caller has to describe what the defensive player should do. To try it out, call (530) 203-8359. While my son is obsessed with baseball, I’m equally obsessed with Voice AI. Voice AI is eating the world. We’re seeing Voice AI consume pre-existing call volume (e.g., lead qualification calls, customer surveys, medical appointment setting) and creating a deluge of new call volume enabled by the Voice AI medium (e.g., role playing for professional development ( https://solidroad.com/ ), an AI that introduces you to people with similar interests ( https://www.boardy.ai/ ), and, yes, voice-first baseball games to make you feel like a better dad). If you’re building in Voice AI and are as enthralled with Voice AI as me, check out what we’re building ( https://ift.tt/M0mpTOJ ) when we’re not building fun voice-enabled AI applications. We love meeting other people in Voice AI and helping them improve their Voice AI agents! Tom https://ift.tt/PsTG8tn https://ift.tt/gKL7Czb October 30, 2024 at 02:13AM
Show HN: Query standup data in natural language Hey HN, I was a PM for ~ 4 years before deciding to ditch everything to learn how to build software on my own. As a PM, I was doing standups every day with our engineers, but whenever I'd need an update I'd still end up pinging them on Slack. Why? - No notes from those standup calls - Folks being unavailable - No solid insight from those standups. We ended up using a Standup tool but it didn't do much apart from converting our calls to text + making standups async. Sadly, there was still no insight. So decided to do build something myself: howsthisgoing - it's an AI-powered standup bot that lets you query team updates using natural language (built with Gemini, Claude, Python/Django, Celery, Tailwind). Demo: https://www.youtube.com/watch?v=9ijCTM3PmaU Technical details: - Set up slack apis for running standups - Use gemini for standup based summaries - embeddings using voyageai - natural language queries handled in 2-parts (first with claude for extraction, then gemini for filtering) - first time setting up RAG pipeline was fun! (let me know if you'd like to know how I did this) Currently exploring integrations with Github/Linear for deeper context on technical updates. Particularly interested in solving the challenge of connecting different data sources (commits, tickets, standups, sales updates) into a queryable knowledge base. Looking for feedback on: - Which data sources would be most valuable to integrate? - Planning to pivot to an "all-in-one updates app with github/linear/hubspot etc. being as source. What do you guys think of this idea? https://ift.tt/mR3oVau October 29, 2024 at 02:20AM
Show HN: Trench – Open-source analytics infrastructure Hey HN! I want to share a new open source project I've been working on called Trench ( https://trench.dev ). It's open source analytics infrastructure for tracking events, page views, and identifying users, and it's built on top of ClickHouse and Kafka. https://ift.tt/yKsf94J I built Trench because the Postgres table we used for tracking events at our startup ( http://frigade.com/ ) was getting expensive and becoming a performance bottleneck as we scaled to millions of end users. Many companies run into the same problem as us (e.g. Stripe, Heroku: https://ift.tt/wvHzncM ). They often start by adding a basic events table to their relational database, which works at first, but can become an issue as the application scales. It’s usually the biggest table in the database, the slowest one to query, and the longest one to back up. With Trench, we’ve put together a single Docker image that gives you a production-ready tracking event table built for scale and speed. When we migrated our tracking table from Postgres to Trench, we saw a 42% reduction in cost to serve on our primary Postgres cluster and all lag spikes from autoscaling under high traffic were eliminated. Here are some of the core features: * Fully compliant with the Segment tracking spec e.g. track(), identify(), group(), etc. * Can handle thousands of events per second on a single node * Query tracking data in real-time with read-after-write guarantees * Send data anywhere with throttled and batched webhooks * Single production-ready docker image. No need to manage and roll your own Kafka/ClickHouse/Nodejs/etc. * Easily plugs into any cloud hosted ClickHouse and Kafka solutions e.g. ClickHouse Cloud, Confluent Trench can be used for a range of use cases. Here are some possibilities: 1. Real-Time Monitoring and Alerting: Set up real-time alerts and monitoring for your services by tracking custom events like errors, usage spikes, or specific user actions and sending that data anywhere with Trench’s webhooks 2. Event Replay and Debugging: Capture all user interactions in real-time for event replay 3. A/B Testing Platform: Capture events from different users and groups in real time. Segment users by querying in real time and serve the right experiences to the right users 4. Product Analytics for SaaS Applications: Embed Trench into your existing SaaS product to power user audit logs or tracking scripts on your end-users’ websites 5. Build a custom RAG model: Easily query event data and give users answers in real-time. LLMs are really good at writing SQL The project is open-source and MIT-licensed. If there’s interest, we’re thinking about adding support for Elastic Search, direct data integrations (e.g. Redshift, S3, etc.), and an admin interface for creating queries, webhooks, etc. Have you experienced the same issues with your events tables? I'd love to hear what HN thinks about the project. https://ift.tt/reuZBTa October 25, 2024 at 07:37PM
Show HN: Building a financial instrument library with Python and Pydantic finstruments is a Python library designed for modeling financial instruments. It comes with the core financial instruments, such as forwards and options, out of the box, as well as position, trade, and portfolio models. finstruments comes with the basic building blocks, making it easy to extend and build new instruments for any asset class. These building blocks also provide the functionality to serialize and deserialize to and from JSON, enabling the ability to store a serialized format in a document database. This library is ideal for quantitative researchers, traders, and developers who need a streamlined way to build and interact with financial instruments. https://ift.tt/xd8G0fX October 28, 2024 at 10:48AM
Show HN: I built EstateLead AI – lead generator boosting agent productivity 10x As a former real estate agent, I discovered a critical pain point in early 2023. Every day, I spent 3-4 hours manually copying leads from Zillow - phone numbers, emails, and property details. This tedious process was stealing valuable time I could have spent with clients. I realized this wasn't just my problem. Thousands of agents were stuck in this same time-consuming routine, manually gathering leads one by one. There had to be a better way. Over the next a months, I faced numerous challenges building EstateLead AI: - Ensuring reliable data extraction - Maintaining compliance with data privacy regulations - Creating an intuitive interface for non-technical users - Scaling the system to handle thousands of requests After countless iterations and feedback from 50+ beta testers, I'm excited to launch EstateLead AI with these key features: Key Features: - One-click lead extraction from Zillow listings - Up to 18,000 leads per month with phone numbers - Export to Excel/CSV for easy CRM integration - Coverage for both US and Canadian markets - Built-in compliance with data protection rules - Simple browser extension interface The tool has already helped 600+ agents save 40+ hours per week on lead generation. Early users report a 10x increase in their prospecting efficiency. I'd love for the HN community to try it out and share your thoughts. What features would make this even more valuable for real estate professionals? Try it here: https://ift.tt/2L9BrnR P.S. First 100 users get 40% off (19 spots remaining), and we offer a 60-day money-back guarantee. https://ift.tt/2L9BrnR October 27, 2024 at 03:02PM
Show HN: Mdx – Execute your Markdown code blocks, now in Go Hey HN! I recently came across makedown here on HN and loved the concept. Wanting to learn Go, I thought this could be a great starter project - so I started working on my own Go implementation, which I’m calling mdx ( https://ift.tt/LsXvBhT ). Key Features: - Define dependencies between commands - Supports shebangs - Ability to pass arguments to code blocks Would love feedback and thoughts! Ref. makedown: https://ift.tt/7QioevR . Thanks for the idea! :) https://ift.tt/LsXvBhT October 26, 2024 at 11:04AM
Show HN: Jist searches JSON 3x faster than jq Hi HN, jist is a JSON path search tool that feels like accessing a JavaScript object. It’s 3x faster than jq for the same use case, and over 2000x more memory efficient when searching a 3.3GB file for the last element (jist reads data in chunks so has constant space use). It doesn’t implement pattern matching yet. A feature I’m looking to add is generating a full JSON schema using the same streaming approach so stay tuned! Any feedback and feature requests are more than welcome https://ift.tt/7SwKHVt October 26, 2024 at 06:55PM
Show HN: Create Waitlist Pages I decided to launch FastWaitlist because I was facing a problem with validating my own product ideas. I noticed the lack of tools that: 1. Let you create waitlist pages fast (always faced complicated page builders) 2. Have conversion optimized templates (we have 5 and we're planning to add more) 3. Support custom domains That's exactly why I launched FastWaitlist. It has solved a huge pain point for me and I hope it'll do the same for you. https://ift.tt/mtV75Xq October 25, 2024 at 11:47PM
Show HN: Infinite SCP Generator I work at .txt, the Outlines maintainers, and did this as a fun side project. I used Outlines to structure language model output in the format of an SCP entry. SCP entries have a specific structure. It's easy to get a language model to generate these structures via JSON without fail -- no parsing errors, no trailing commas, no missing brackets. Outlines ensures that all text generated will follow the SCP structure. We're also testing an internal structured text API, and figured this might be a good way to test out repeated, regular calls. Entries are generated at midnight and uploaded to a simple website. Check out the resulting entries here: https://ift.tt/vHMcpKS https://ift.tt/3WZmwi0 October 24, 2024 at 11:46PM
Show HN: Hyper Workspace – A no-code framework for JSON Hello HN ! I'm Florian, and I would like to share with you a project I've been working on over the past few months: Hyper Workspace. Before working in tech, I have been in the finance industry for more than 10 years. During this time, I've experienced using Excel+VBA and Python notebooks as main interfaces, but never felt completely satisfied. Interfacing potentially complex and nested objects that you need to consume in your applications is still a pain. Modern browsers can offer much more in terms of UI / UX, especially when coupled with JSON data representation, but they require to write and deploy specific code, use a rendering library... and I believe there is a better way: JSON deserves a fully horizontal framework, abstracting web semantics, and filling the gap between data, views and behavior. To this end, I've been trying to shape Hyper Workspace so that it can provide no-code tools to build Documents': - Fields: strongly typed, deeply nested and potentially recursive, data structures (inspired by Typescript). - Views: drag n' drop configuration of documents' pages (ie sub-objects): layout, fields, widgets, actions, graphs. - Methods: visual configuration of your "business logic" using a Domain Specific Language, plugged at every layer of your data structure, with an "object-oriented programming" philosophy. Builtin "Modules" can be used to add more functionalities. You can build "standalone" Documents (like spreadsheets) but also Collections where Records share the same Schema, like a NoSQL version of Airtable. Unstructured data (pdf, images, audio, video) are stored as "Files" that can be linked inside Documents. All this in a multi-desktops OS-looking environment. You will find a complete documentation inside your workspace. So far, this is a purely local application that stores everything in your IndexedDB. It can be installed locally (PWA) and will then work completely offline. No account required and nothing leaves your device. This is more the early stage of a framework than a complete product, but I hope it can shed a light on a potential new way to "configure" interfaces or even workflows, given that Documents can be seen as stateful representations of business processes. Builtin "Modules" are the key to extending the framework's capabilities: calling APIs, using AI agents to structure or query documents, or even take action... I'd love to read your feedbacks, questions and ideas on it, and see together what could make it useful for you ! Thanks ! Florian https://ift.tt/n0Gbhgw October 22, 2024 at 11:04PM
Show HN: Data Formulator – AI-powered data visualization from Microsoft Research Creating data visualizations with AI nowadays often means chat, chat and more chats...and writing long prompts can be annoying while they are also not the most effective way to describe your visualization designs. Data Formulator blends UI interaction with natural language so that you can create visualizations with AI much more effectively! You can: * create rich visualizations beyond initial datasets, where AI helps transforming and visualizing data along the way * iterate your designs and dive deeper using data threads, a new way to manage your conversation with AI. Here is a demo video: https://ift.tt/Yf3D6oG.... Give it a shot and let us know how it looks like! https://ift.tt/bwMF4tI October 22, 2024 at 01:12AM
Show HN: Llama Workspace – An Open Source ChatGPT Teams Alternative Hey HN! We are happy to share with you our latest project: Llama Workspace ( https://ift.tt/WylG3j9 ). It's an open-source ChatGPT Teams alternative made with the needs of businesses and organizations in mind. It has everything you can expect from ChatGPT, plus first class support for users and roles management, as well as advanced collaboration features. We've made it very easy to self-host anywhere you like, and we also provide a cloud version. So no excuse for trying it out! Why using Llama Workspace instead of ChatGPT Teams? - You have access to all the major Large Language models in one place, including GPT-4, Claude Sonnet or Gemini. - It brings savings of about 70% - 85% when compared to ChatGPT Teams (see website for more info on this). - You can integrate it with your own code, like AI agents, and provide a single platform to access AI. Here's the Github link: https://ift.tt/z5AOw7D I would love to hear your ideas, experience, and feedback about the product! What should we implement next? https://ift.tt/WylG3j9 October 21, 2024 at 08:08PM
Show HN: Floating point arithmetic types in C++ for any size and any base Since there is so much interest on HN in floats lately and their software implementations, I wanted to show mine. It has no use and is just for teaching me floats and C++. Give me your thoughts. https://ift.tt/OMZIFxU October 18, 2024 at 11:12PM
Show HN: Launching GetProduct.dev I would be happy to get any feedback! Here is why I built it and what it is: I found it hard to monetize another project of mine where users entered wishes they had. Google ads generated almost no revenue. Therefore I tried Affiliate Marketing using various Affiliate Programs where the users wish is entered as query. Especially with amazon this worked pretty well even though I hit a lot of problems to get it working. My revenue from that is at least 10x the revenue I had from google ads. Thats why I build getProduct.dev and I would be really happy if you would try it our and give me some feedback. It is a nicely documented api to get dynamic affiliate links. I do not get a cut of the revenue because it uses your own Amazon account. You get 100 free requests once you create an account and if you need more for testing you can contact me. Currently it only supports amazon affiliate but I plan on integrating more like Aliexpress and also I am thinking about more usecases that could potentially leverage AI to generate even better responses. Thanks for reading and any feedback! https://getproduct.dev/ October 21, 2024 at 03:12AM
Show HN: Create mind maps to learn new things using AI Enter a topic and get a learning mind map generated by an LLM with links to learn more about each subtopic. You can use it with local models (through Ollama) or external models. If you have any feedback, please share it! Hope it's useful https://ift.tt/ouAs7XV October 21, 2024 at 01:31AM
Show HN: Open-Source Zero-Shot Image Model Server Enabling Model Feedback Hi everyone! Here is an open source implementation of a decently performant server hosting zero-shot image models (CLIP for image classification, OWL-ViT-ST for object detection), with an extra algorithm to allow users to give the models feedback when they make mistakes! We built a company off this flavor of tech two years ago and have clients who are currently using our commercial API. We are now moving on to other projects but want to make sure our clients still have access to the approaches that they've grown to rely on, so we're open sourcing a simple implementation that they'll be able to use after we've shut down our hosted API! I used to work at a robotics startup. After a while it seemed clear that the biggest limiting factor in our ability to ship new models wasn't innovation on model architecture, it was access to relevant, high-quality training data. Around that time CLIP was released, which got me thinking about the idea of having models with world-knowledge baked in so as to reduce the amount of training data required. A year later when Stable Diffusion dropped, my cofounder Ben Brooks and I took the plunge and founded DirectAI, where we worked on building ways to get performant models without collecting any training data, using the knowledge stored in pretrained models instead. In this implementation, we replace the linear classification head typically used in zero-shot image classifiers with a modified nearest neighbors method that lets you use multiple examples (both positive and negative) per-class to make sure the decision boundary the model is using is more aligned with what you had in mind. Our clients have found it very useful for things from interior design to content moderation to sports analytics, building models that are either too niche to be supported by a traditional cloud-hosted computer vision API or are subtly different from the models that existing cloud APIs host. For example, one of our clients wants to filter out all images containing alcohol. Hive has an API for that, but Hive explicitly allows red solo cups that don't obviously have anything alcoholic in them, whereas our client wanted to filter those out too! Feedback is welcome! There are still bugs in the Gradio frontend / codebase in general, but I have a deadline and need to be working on new stuff at a new job starting Monday so I thought I would just go ahead and get it out there! I've never tried to publish a real open source piece of code before and I must admit I am quite nervous! https://ift.tt/G3Mnmyl October 20, 2024 at 12:21AM
Show HN: I made a site to quick identify any plant and learn how to care for it I'm still working on it, curious but what feature may be most valuable for you? What do you think about personalised care instructions, and an interactive chat feature for each plant. https://frondly.app/ October 20, 2024 at 02:30AM
Show HN: Contagious Beliefs–Simulating Political Alignment This is a simulation that allows you to test how beliefs are adopted based on pre-existing beliefs and how they spread through a population. It is entirely editable, so you can test your own parameters based on your own data or intuitions. A detail methodology can be found here: https://ift.tt/Gy4h6zd... https://ift.tt/gTfOs4w October 20, 2024 at 12:23AM
Show HN: A browser-based x86-64 assembly playground I recently created this project to teach assembly, and to have an assembly playground in the style of codepen/jsfillde to embed in my blog posts. It's based on a wasm version of the blink x86-64 emulator, so everything runs client side with no backend server and with a runtime of only ~200Kb https://x64.halb.it/ October 18, 2024 at 11:02PM
Show HN: Thread Archive – free tool to save Slack community data Slack's new content deletion policy is a nightmare for 1000s of communities on Slack. It means that all messages older than one year will be permanently deleted from their servers. Thread Archive gives you a simple way to save all your Slack community's chats in an accessible archive. https://ift.tt/UdrKtqa October 18, 2024 at 03:30AM
Show HN: React-guitar-chord, a react component for chord name to guitar diagram I made a pretty simple react component that takes in a chord name and spits out a guitar chord diagram. So far it has no regard for playability or bar chords or any alternate positions, and doesn't ensure that all the notes of a complex chord are played. It has a bias towards open notes so with enough alterations it tends toward an open strum. Still, it's pretty accurate for simpler 3 or 4 note chords. I think my next step will be ensuring that 4 or less fretted notes are included in a chord diagram, using bar chords and moving up the neck if needed. https://ift.tt/fiv1KMV October 18, 2024 at 12:29AM
Show HN: I made a webapp where you can share your projects Hi guys I made this simple web application, you can use it to share your projects. Your project will appear only if you visit the projects of the other users. In the 'you' section, add an app ( name, link, [icon_link]) and apply the starting points you got, to the app so that all the other N users will see it (where N indicates how many points you applied to the app). IT'S TOTALLY FREE NO PASSWORD NEEDED FOR LOGIN https://appearnow.me October 17, 2024 at 10:18PM
Show HN: Screvi – Stop Forgetting What You Read If you’re anything like me, you read and highlight a lot but always seem to forget most of it. Not only that, but I find myself spending too much time scrolling Twitter and Instagram these days. I needed an app with similar features, but one that would make me productive instead of feeding me addicting and pointless content. The usual notion/obsidian workflows are nice, but they get overly complex quick and I end up wasting a lot of time on them. I needed something easy and accessible like a social media app. So, I built an app that fixes these issues by capturing and transforming my thousands of highlights into a scrollable, Instagram-like feed. key features - Sync highlights from Kindle - Capture physical book highlights - Save from Twitter - Transcribe audio notes (so I can add highlights/thoughts while running or walking) - Feed/Stories system Let me know what you think! And if you find any bugs or have any ideas, please let me know at support@screvi.com https://screvi.com October 17, 2024 at 01:23AM
Show HN: I made a social like Instagram but for books Hey Book Lovers! Today, we’re thrilled to introduce Biblio.com, your new social playground for everything books! Whether you're a voracious reader, budding author, or just someone who loves to explore new ideas through the written word, Biblio has it all. With Biblio, you can: Upload your own stories, novels, or poems and share them with the world Read an ever-growing library of books and writings by other users – discover hidden gems! Follow authors, readers, and creators whose content you love – build your literary network Create reading lists, leave reviews, and join discussions in vibrant community forums Get personalized recommendations and discover your next favorite book in no time! Why Biblio? Unlike other book apps, Biblio is truly multi-functional – it’s your personal bookshelf, a space to publish your work, a place to connect with readers and writers, and much more. Forget juggling between multiple platforms; now you can do it all in one spot! Whether you're an indie author looking to build a fanbase or a casual reader searching for your next cozy read, Biblio makes it effortless to connect and share stories. We’re just getting started and would love your feedback! Jump in, explore, and help us shape the future of books online. Join the conversation – the literary world is waiting! https://www.bibliou.com October 16, 2024 at 01:21AM
Show HN: 4B+ DNS Records Dataset Hi HN, I've been working on building a pipeline to create a DNS records database lately. The goal is to enable research as well as competitive landscape analysis on the internet. The dataset for now spans around 3 billion records and covers all the common DNS record types: A AAAA ANAME CAA CNAME HINFO HTTPS MX NAPTR NS PTR SOA SRV SSHFP SVCB TLSA TXT Each line in the CSV file represents a single DNS record in the following format: https://ift.tt/x8XJNPv I believe this dataset will be valuable for various research purposes and competitive analysis. Let me know if you have any questions or feedback! https://ift.tt/ERXJBgq October 16, 2024 at 12:56AM
Show HN: Turn any YouTube video into a summary, quiz, & flashcards YouTube is the greatest learning resource in the world. Yet it doesn't provide any tools to help you learn from videos. I built a simple app that turns YouTube videos into structured learning material using AI. It generates a summary, quiz and flashcards from video content, meaning you can learn from anything (well, at least any YouTube video). Try it out. Would be happy to keep building features based on feedback! https://asterlab.io/ October 14, 2024 at 11:15PM
Show HN: Those Who Die as Cattle What if a game gave you the backstory of everyone you killed or got killed? I feel like I've heard this question asked before, but I've never seen it done. For Ludum Dare 56 (a game jam), I spent 48 hours building a prototype called Those Who Die As Cattle, a WW1 trench warfare simulation game. I used ChatGPT 4o-mini to generate details for the 12,000 soldiers involved in the game. Notably, I had to supply names because it couldn't introduce sufficient variety. I also had to add randomly generated adjectives for the soldiers; otherwise, the prompts were too similar, even with different seeds and temperatures. Initially, I used Claude (which I usually prefer), but it made every soldier a former carpenter’s apprentice. Since this is a game and your actions should matter, instead of generating one block of text, I had it generate JSON by providing a template with cues for different things that might happen to the soldier. I used these to dynamically assemble a story for each casualty. For example: { "briefBio": "Tell me a little bit about the solider", // Putting this first is critical to introduce variety … "lastThoughtsArtilleryDead": "What are his last thoughts as he dies from an artillery strike" , … "effectOnOthersShotInjured": "What are the long term impacts on the people who care about him when he is seriously injured by a bullet wound", } The prompt template can be found here: https://ift.tt/7LvMzJ6... The underlying simulation was programmed using Unity ECS (Entity Component System) and simulated, to some degree, each soldier and artillery piece. Casualties were taken in entity order, so there's some predictability regarding when each name gets added to the list, though it isn't completely deterministic. The area where I skimped the most was the UI, but the game is in a playable state if you're willing to look at the center column and figure out what the numbers mean. A final notable learning is that Unity UI toolkit (their modern, web dev like solution) performs significantly better on the web than in the native linux build, at least on the 2022 LTS version of Unity. This suggests that they may be leveraging the rendering efficiencies built into browsers. https://ift.tt/QGTjyNK October 14, 2024 at 09:32PM
Show HN: Book to Classical Music Soundtrack Hey, HN, literature and classical music are two of my biggest passions, and I always found it interesting how some classical pieces fit to the stories they are barely connected with. E.g. some of Tchaikovsky's music will always evoke the world of Harry Potter for me; music of Sibelius takes me to the same landscapes as writings of Tolkien - of course, these clearly result from John Williams' and Howard Shore's direct inspirations from the romantic composers when creating the soundtracks for the movie adaptations. For a more niche example, music of Olivier Messiaen fits in my view the book L'Écume Des Jours by Boris Vian - grandiose movements of the Turangalila symphony for the first half of the story, the expressionism and sadness of Vingt Regards the latter. The examples are numerous. My recent idea was to somehow automate this pairing of literature and music. It works quite simply - at first, I collected a large dataset of classical music pieces with text descriptions, which I transformed into embeddings. When generating a soundtrack, I use an LLM to create a description of what kind of music would be suitable for the story, and the embedding of this description is used to fetch the most relevant tracks. While the resulting playlist can barely be understood as "soundtrack" at this point, it can be a pleasant background playlist for reading the book. Thanks in advance for any suggestions or feedback :) https://ift.tt/mK03tps October 14, 2024 at 02:58AM
Show HN: SSO – Small String Optimization I wrote this just for fun when saw article about sso in Rust[1]. My string can store up to 23 (excluding null-terminator) 8-bit chars without calling allocator. Here I can mistake, but.. Curious fact: both - libstdc++[2] and libc++[3] - do access to union member without any check that it is active now. AFAIK, this is UB in C++. But I assume that they just rely on theirs compiler features. I tried to avoid this using `std::byte[]`. But I'm still sure that there are several UB's in my code :) [1] https://ift.tt/AhJrbfi... [2] https://ift.tt/fgRO29v... [3] https://ift.tt/ZWKSfO2... https://ift.tt/rhLnfMX October 12, 2024 at 11:56PM
Show HN: ZenMD converts a Markdown folder to a site with a single command I have been dogfooding this npx tool to build all my markdown content sites. It allows me to start a site within minutes. As it's just `npx zenmd && netlify deploy` without downloading any framework starter kit, cloning any repo, or generating any folder structure with configs. It's built on remark.js, supports images, [[wikilinks]], and raw html in markdown, as well as custom layout.html customizations. Once the site is built, you can open the output `.html` pages without a server to preview. Demo: here is a post expanding on the "whys" on my personal site (built with ZenMD) - https://ift.tt/qdY7pUo Just made the Github repo: https://ift.tt/X9xd6F8 public if you'd like to look at the code. https://ift.tt/zvKGlmr October 11, 2024 at 09:54PM
Show HN: QuickDessert – generate quick dessert recipes and adjust Ingredients It’s not just your average recipe site – they’ve got an AI-powered feature that generates custom dessert recipes based on your preferences or available ingredients! If you’re short on time or ingredients, the site lets you tweak recipes on the fly. You can adjust ingredients to suit your pantry, dietary preferences, or time constraints, and the AI will instantly generate a new dessert idea for you. Super handy for those moments when you have a sweet tooth but not everything in your kitchen! For example, I tried out a brownie recipe and swapped out sugar for honey – worked like a charm! The flexibility and ease make it really fun to experiment with desserts without spending hours searching for ideas or second-guessing ingredient swaps. https://ift.tt/MRj0dal October 9, 2024 at 07:51PM
Show HN: Build Apps via Voice TiramAi - Beta - We have built an app that is now available both on App Store & Play Store that allows non-technical users to build web apps, mobile apps and logic apps as it seems fit by generating user stories and personas. Can be refined as many times. For web apps, the code gets generated in html, cs, JavaScript for the frontend and uses firebase as the db. For mobile app, flutter code gets generated and it takes us a day or so to get it out into test flight or beta testing, due to the app distribution hurdle. Logic apps use c# minimal api and gets deployed on Azure / Google. Exporting of User stories is behind a paywall. We are still fine tuning the code output for all the apps. You can provide feedback within the app as well by shaking the device. Any input / feedback / thoughts are appreciated. https://www.tiram.ai October 11, 2024 at 04:08AM
Show HN: Arch – an intelligent prompt gateway built on Envoy Hi HN! My name is Adil Hafeez, and I am the Co-Founder at Katanemo and the lead developer behind Arch - an open source project for developers to build faster, generative AI apps. Previously I worked on Envoy at Lyft. Engineered with purpose-built LLMs, Arch handles the critical but undifferentiated tasks related to the handling and processing of prompts, including detecting and rejecting jailbreak attempts, intelligently calling “backend” APIs to fulfill the user’s request represented in a prompt, routing to and offering disaster recovery between upstream LLMs, and managing the observability of prompts and LLM interactions in a centralized way - all outside business logic. Here are some additional key details of the project, * Built on top of Envoy and is written in rust. It runs alongside application servers, and uses Envoy's proven HTTP management and scalability features to handle traffic related to prompts and LLMs. * Function calling for fast agentic and RAG apps. Engineered with purpose-built fast LLMs to handle fast, cost-effective, and accurate prompt-based tasks like function/API calling, and parameter extraction from prompts. * Prompt guardrails to prevent jailbreak attempts and ensure safe user interactions without writing a single line of code. * Manages LLM calls, offering smart retries, automatic cutover, and resilient upstream connections for continuous availability. * Uses the W3C Trace Context standard to enable complete request tracing across applications, ensuring compatibility with observability tools, and provides metrics to monitor latency, token usage, and error rates, helping optimize AI application performance. This is our first release, and would love to build alongside the community. We are just getting started on reinventing what we could do at the networking layer for prompts. Do check it out on GitHub at https://ift.tt/d2MUo9S . Please leave a comment or feedback here and I will be happy to answer. (I did make a mistake earlier not properly tagging the title with Show HN and the post didn't end up in Show HN. Later I updated the title but then my text started showing up as a comment. So I redid the post again, sorry if you had looked at the previous post already. Here is the link to the old post https://ift.tt/pUoFdjY ) https://ift.tt/w48Wcq5 October 10, 2024 at 11:15PM
Show HN: VRE Dataset generation for MultiTask vision models training from videos Been working on this tool for my PhD which involves training multi task vision models using various pre-trained models as inputs or pseudolabels in order to improve generalization. I work mostly on UAV datasets, but it should work okay on indoor scenes or self driving (at least Marigold and Mask2Former). For example, this dataset was generated using this tool: https://ift.tt/aTtnFWQ I'm quite aggressively trying to "just get the nn.Module" from the public repos that other researchers put up in their overly convoluted frameworks. A simple `forward(rgb_input: torch.Tensor) -> torch.Tensor` is nice, having 100 imports from a generic framework that has versions incompatibilities with everything else is not. PS: most mains are standalone runnable too, i.e. - https://ift.tt/ObdmBoC... or - https://ift.tt/ObdmBoC... https://ift.tt/C4VQ58I October 9, 2024 at 11:09PM
Show HN: Actor Kit – XState Machines for Cloudflare Workers Actor Kit is a library for running state machines in Cloudflare Workers, leveraging XState for robust state management. It provides a framework for managing the logic, lifecycle, persistence, synchronization, and access control of actors in a distributed environment. Key features include: - Persistent, distributed state management - Type-safe events and state with TypeScript and Zod - Real-time updates and seamless server-to-client state transfer - Access control and secure data management - Framework-agnostic with examples for Next.js and Remix integration - Actor lifecycle management in distributed environments - Support for real-time collaboration - State migration support The project is in early development, and I'm seeking feedback on the ideas and APIs. Cheers, Jon https://ift.tt/mOt0TM6 October 8, 2024 at 11:56PM
Show HN: Running Game Boy ROMs on the STM32 ARM Cortex Microcontroller Got Tetris running on a 26€ STM32F429 Discovery board emulating the Game Boy CPU and graphics on the ARM Cortex M4. Really fun project to hack together, need to try the original Game Boy Pokemon (red/blue) games next. https://ift.tt/zgFZue7 October 8, 2024 at 10:33AM
Show HN: I made it even easier to self-host my open-source form back end On a bit of a mission here. Just made the first addition to what I plan to be lots of variations of self-hosting this Formspree alternative. The goal is to allow you to host just what you need, and avoid the bloat and complexity of self-hosting the entire service. https://ift.tt/O3LYuwW October 6, 2024 at 11:10PM
Show HN: CodeAir – Control VS Code from Your Phone Hey HN, Inspired by the classic "Compiling" XKCD comic ( https://ift.tt/aSUWQg4 ), I’ve built CodeAir (codeair.app) – an Android app and VS Code extension that turns your phone into a remote control for VS Code. With CodeAir, you can: - Control 800+ VS Code commands right from your phone. - Send text and image files directly into your workspace and even have the images transcribed using AI. So imagine taking a snapshot of a piece of code from a physical textbook and being directly able to run that inside VS Code. - Request screenshots from VS Code while AFK and get them delivered to your phone so you can monitor long builds or processes remotely when you're on a coffee break. - If you’ve got an OpenAI API key, you can even perform QnA over your transferred files and images. I made this tool to help monitor long builds or tests when I’m AFK, but it’s useful for so much more! It’s super niche, but I hope some of you find it helpful. Check it out at codeair.app and let me know what you think. Feedback is very welcome! https://www.codeair.app October 5, 2024 at 06:35PM
Show HN: Production Quality AI True Crime Podcast Generator Hello HN! My brother and I built Golpo, the first production-quality AI true crime podcast generator. Just tell Golpo what you want to listen to and it creates a real true crime podcast (usually > 30 min long) in minutes. Golpo does everything that a human podcastor would do. It first understands what you want to listen to, does highly detailed research on the case, creates a script and then uses that to create a production-quality podcast Golpo ensures that every story is meticulously fact-checked and based on real-life events. Say goodbye to waiting for new episodes and experience true crime made just for you on demand. Available today. No waitlist! Web: golpo.app App Store: https://ift.tt/nG5Ksxo... https://golpo.app October 6, 2024 at 02:30AM
Show HN: open source framework OpenAI uses for Advanced Voice Hey HN, we've been working with OpenAI for the past few months on the new Realtime API. The goal is to give everyone access to the same stack that underpins Advanced Voice in the ChatGPT app. Under the hood it works like this: - A user's speech is captured by a LiveKit client SDK in the ChatGPT app - Their speech is streamed using WebRTC to OpenAI’s voice agent - The agent relays the speech prompt over websocket to GPT-4o - GPT-4o runs inference and streams speech packets (over websocket) back to the agent - The agent relays generated speech using WebRTC back to the user’s device The Realtime API that OpenAI launched is the websocket interface to GPT-4o. This backend framework covers the voice agent portion. Besides having additional logic like function calling, the agent fundamentally proxies WebRTC to websocket. The reason for this is because websocket isn’t the best choice for client-server communication. The vast majority of packet loss occurs between a server and client device and websocket doesn’t provide programmatic control or intervention in lossy network environments like WiFi or cellular. Packet loss leads to higher latency and choppy or garbled audio. https://ift.tt/3vlLIA6 October 4, 2024 at 10:31PM
Show HN: I made a super-simple image CDN Hi HN, MageCDN is a simple and affordable image hosting service I have been working on for the past few weeks. The idea came out of my own frustration with hosting and optimizing images for my blog. While platforms like Imgur make it really simple to upload images, they don't allow you to embed them. Services like Cloudinary, Imagekit exist, but I found them too complex for my needs. Plus, they get really expensive past their free tier. So, I started MageCDN with three simple goals: - pricing should be affordable and scale linearly. - basic image operations (resize, crop, optimize) should be doable within the app. - uploading and getting a link you can use should be fast and hassle-free. Would love to hear what you think! https://magecdn.com/ October 3, 2024 at 09:04PM
Show HN: I have published a mobile HN client to have coffee chat with HN users Hi, I have published a mobile app to have coffee chat with HN users. There was already a guy who made a website for this, but I wanted to do it in mobile to practice WebRTC(technology behind the audio chat) and have some fun. I haven't spent too much time in terms of the general functionality as a HN client but can upgrade it if there are enough users enjoying the app. You can download the apps via the links below; IOS: https://ift.tt/E0DIRyW... Android: As you might have known, Android requires the app to be tested in a closed beta with at least 20 people So you can follow these steps to be a part of it and download; 1 - join the group https://ift.tt/LPSMDGe 2 - download the app and enjoy https://ift.tt/zKLxqMB... or https://ift.tt/6AFIxOM Feel free to contact me if you have any questions; https://ift.tt/ZQKAcR4 October 3, 2024 at 09:43PM
Show HN: My Voice AI interviews you, then writes your LI / X posts I'm Tom, the cofounder of Canonical AI. I build a Voice AI agent that interviews you, then writes LinkedIn and X posts based on the interview. ### Why did I build this? 1. LinkedIn and X are b2b lead goldmines. But generating content is hard. You have to figure out what to say. If you're like me, you have trouble coming up with social post ideas when staring at a blank screen. But interesting ideas arrive when I step away from my laptop and talk to people. Also, it's hard to compose the content in a way that appeases the platform algorithms. 2. I want to dogfood my product. We’re building Mixpanel for Voice AI Agents ( https://ift.tt/zmOdraC ). We map the caller journeys. We provide audio metrics like latency and interruptions. We show Voice AI devs why and when their agents are failing. I want a Voice AI agent so I can be a user of my Voice AI analytics product. 3. Voice AI will eat the world! This project is essentially a Voice UI for a social media content generation platform. We’re seeing more and more Voice AI interfaces like this (e.g., NotebookLM converts any content into a podcast), and many more use cases for Voice AI agents. ### How did I build this? I built the Voice AI agent using Vapi ( https://vapi.ai/ ). Vapi is amazing! It’s by far the best end-to-end Voice AI platform out there. I actually had to add latency to the Voice AI. I am a slow thinker / talker and it was unrealistic that _my_ Voice AI clone would respond at normal human intervals. I cloned my voice using PlayHT ( https://play.ht/ ). Yes, my voice really is that airy. I wish I could say it’s an issue with PlayHT’s voice cloning, but it’s spot on. The hardest part has been making the social posts feel like they weren't written by an LLM. I initially was using 4o-mini, but they definitely felt written by...well...GPT. I broke the steps into smaller pieces, making one LLM call that extracts the key insights and quotes and another LLM call that generates the content. That helped, but the social posts still felt LLMy. Then I switched to using Claude. That made a big difference. I love Claude. I use Claude so much I'm not sure where I end and Claude begins. I'd love to hear suggestions from the community on how to make the LLM-generated content feel less, as a friend of mine said, boring AF. ### Example Output LinkedIn: Did you know that most Voice AI agent developers are still manually listening to call recordings? It's a time-consuming and ineffective approach to knowing what your agent is doing. It's like drinking from a fire hose. At Canonical AI, we’re changing that by providing Voice AI developers with caller journey maps, audio metrics, and conversation analytics. X: Voice AI developers are currently flying blind. At Canonical AI, we're changing that by building Mixpanel for Voice AI agents. You can see more on my recent posts on my LinkedIn ( https://ift.tt/jrLfp2e ) and Twitter ( https://ift.tt/MAo6BX0 ). ### Try it out! Here's the number: +15302038359 October 3, 2024 at 02:38AM
Show HN: Hexabot, Open-source AI Chatbot/Agent Tool Why Hexabot? - Fully open source (AGPLv3) - Easy-to-use interface - Multilingual and multi-channel support - Seamless integration with existing workflows and Text-to-Action Join the growing community of developers and businesses around the world building with Hexabot! Check it out: https://ift.tt/lPDXQqh https://ift.tt/k908FSL October 3, 2024 at 01:32AM
Show HN: Kameo – a Rust library for building fault-tolerant, async actors Hi HN, I’m excited to share Kameo, a lightweight Rust library that helps you build fault-tolerant, distributed, and asynchronous actors. If you're working on distributed systems, microservices, or real-time applications, Kameo offers a simple yet powerful API for handling concurrency, panic recovery, and remote messaging between nodes. Key Features: - Async Rust: Each actor runs as a separate Tokio task, making concurrency management simple. - Remote Messaging: Seamlessly send messages to actors across different nodes. - Supervision and Fault Tolerance: Create self-healing systems with actor hierarchies. - Backpressure Support: Supports bounded and unbounded mpsc messaging. I built Kameo because I wanted a more intuitive, scalable solution for distributed Rust applications. I’d love feedback from the HN community and contributions from anyone interested in Rust and actor-based systems. Check out the project on GitHub: https://ift.tt/cTbqwXI Looking forward to hearing your thoughts! https://ift.tt/cTbqwXI October 2, 2024 at 11:52PM
Show HN: NewTabTodo Firefox Extension replaces your new tab with todo-list Hello HN, As an avid Firefox user, I have built a simple extension that replaces your new tab with a to-do list. You can add or delete your to-dos without leaving your browser. I felt that an extension like this didn't exist for Firefox, so I decided to build it myself. It features a minimal and clean UI. I am open to feedback Link to code: https://ift.tt/VsBZ1cD https://ift.tt/VdJS0Fi October 1, 2024 at 01:25AM
Show HN: PurePlates – A Recipe Scraping iOS App Hey Everyone! This is my first post on Hacker News, and I wanted to share my project with you! From a young age, I dreamed of becoming a chef, but life took me down the path of software engineering instead. Nevertheless, my passion for creating delicious meals has always remained close to my heart. This journey led me to seek out recipes to enhance my cooking skills, but I often found myself frustrated by the websites I encountered—ads everywhere, walls of text, and recipes that were difficult to follow. That’s where PurePlates comes in. PurePlates allows users to share a recipe through their mobile browser or simply copy and paste the URL directly into the app. In an instant, you receive a scraped recipe that includes instructions, ingredients, and any available nutrition information. You can easily start cooking and follow along step by step, making the process much more enjoyable. Plus, if you find a recipe you love, you can add it to your favorites for quick access later! As this is just an MVP, there’s still plenty of room for improvement, and I’m actively working on enhancing the extraction rate for recipes. Thanks for reading, and keep hacking! Connor https://ift.tt/hfCiHSM September 30, 2024 at 08:59PM
संदेश
