Show HN: Fin.flights 2.0 – AI-powered flight search https://bit.ly/4dmb4Ag

Show HN: Fin.flights 2.0 – AI-powered flight search fin.flights is an AI-powered flight search tool that uses natural language processing to find the best flights based on user queries. just released version 2.0 with significant improvements: - No login required: Start searching immediately without creating an account - Advanced AI: Utilizing latest LLMs for more accurate flight results - Natural language interface: Type queries as you would ask a friend - Global coverage: Searches flights worldwide with multi-language support - Free to use: Only upgrade if you find it valuable How it works: 1. Enter a natural language query (e.g., "cheapest business class flights from NYC to Tokyo in the next 3 months with max 1 stopover") 2. Our AI interprets the query and searches across multiple airlines and booking platforms 3. Results are displayed instantly, sortable by price, duration, or best value I built this because existing flight search engines often require too many inputs and don't offer the flexibility of natural language queries. Myy goal is to make flight search as simple as asking a knowledgeable friend. I'd love to get feedback from the HN community on the user experience, AI accuracy, and any features you'd like to see added. https://bit.ly/4dwVNfY September 29, 2024 at 11:02PM

Show HN: FlowG – A KISS low-code log processing software https://bit.ly/4duvE1g

Show HN: FlowG – A KISS low-code log processing software https://bit.ly/4dsXaMP September 29, 2024 at 04:22AM

Show HN: htmgo - build simple and scalable systems with golang + htmx https://bit.ly/4dvH9Wi

Show HN: htmgo - build simple and scalable systems with golang + htmx Hey all, I just wanted to share a project I've been working on for the past month. After years of heavy frameworks, I really like the idea of using htmx, but it’s a little too low level for me and needs a thin layer above it to facilitate things like components, better syntax with complex JS inside of an attribute, etc To try and solve this problem with a very minimal stack (golang + htmx) that I've been really enjoying, I'm building this project to cater to my needs and was thinking it would be useful for other developers. https://bit.ly/401p7s7 September 28, 2024 at 10:34PM

Show HN: Iceoryx2 – Fast IPC Library for Rust, C++, and C https://bit.ly/47GMY21

Show HN: Iceoryx2 – Fast IPC Library for Rust, C++, and C Hello everyone, Today we released iceoryx2 v0.4! iceoryx2 is a service-based inter-process communication (IPC) library designed to make communication between processes as fast as possible - like Unix domain sockets or message queues, but orders of magnitude faster and easier to use. It also comes with advanced features such as circular buffers, history, event notifications, publish-subscribe messaging, and a decentralized architecture with no need for a broker. For example, if you're working in robotics and need to process frames from a camera across multiple processes, iceoryx2 makes it simple to set that up. Need to retain only the latest three camera images? No problem - circular buffers prevent your memory from overflowing, even if a process is lagging. The history feature ensures you get the last three images immediately after connecting to the camera service, as long as they’re still available. Another great use case is for GUI applications, such as window managers or editors. If you want to support plugins in multiple languages, iceoryx2 allows you to connect processes - perhaps to remotely control your editor or window manager. Best of all, thanks to zero-copy communication, you can transfer gigabytes of data with incredibly low latency. Speaking of latency, on some systems, we've achieved latency below 100ns when sending data between processes - and we haven't even begun serious performance optimizations yet. So, there’s still room for improvement! If you’re in high-frequency trading or any other use case where ultra-low latency matters, iceoryx2 might be just what you need. If you’re curious to learn more about the new features and what’s coming next, check out the full iceoryx2 v0.4 release announcement. Elfenpiff Links: * GitHub: https://bit.ly/4dGdWbv * iceoryx2 v0.4 release announcement: https://bit.ly/47L0nG6 * crates.io: https://bit.ly/3Y1IwHJ * docs.rs: https://bit.ly/47I4ymf https://bit.ly/47L0nG6 September 28, 2024 at 05:40PM

Show HN: Open-source Docker image to run Chrome browsers in your cloud https://bit.ly/3N3NeOR

Show HN: Open-source Docker image to run Chrome browsers in your cloud If you’ve ever worked with web scraping or automation using Puppeteer or Playwright, you know that running Chrome can be tricky. While headless mode (Chrome with no display) often works, some page elements only render in headed mode, forcing you to run Chrome with a display. This usually means bloating your Dockerfile with complex dependencies like xvfb. Updating your server or Lambda container to handle this can increase costs and lead to dependency hell. The Solution: I’ve built a streamlined solution—a Docker image that runs Chrome with all its necessary dependencies and exposes a FastAPI interface. This decouples the browser logic from your main codebase, letting you focus on scraping or automation without worrying about the underlying setup. The container exposes a WebSocket endpoint at ws://localhost:8000/ws, allowing you to connect Playwright or Puppeteer over Chrome DevTools Protocol (CDP). For example: browser = await pw.chromium.connect_over_cdp('ws://localhost:8000/ws') Each connection to the WebSocket spins up a new Chrome session in headed mode, and you can run multiple sessions concurrently. Unlike other solutions, Finic is fully open-source (Apache 2.0 licensed). You can try it by pulling the official Docker image or cloning the repo at Finic on GitHub ( https://bit.ly/3XKPBw6 ) and running `docker-compose up --build`. This approach simplifies handling Chrome for scraping, saves on cloud costs, and keeps your setup clean. Give it a try! https://bit.ly/3ZCDzq8 September 27, 2024 at 04:06PM

Show HN: DecisionBox – Continuous Accuracy Improvement for LLM Apps https://bit.ly/4eqTZGQ

Show HN: DecisionBox – Continuous Accuracy Improvement for LLM Apps Just released DecisionBox, an open-source SDK that helps developers make high-accuracy decisions in LLM apps, which continuously improve with more data. DecisionBox tackles the challenge of maintaining decision accuracy beyond the limitations of prompt engineering. It streamlines the data science process with an easy-to-use API and enables ongoing accuracy metrics and model improvements. Get started: https://bit.ly/3TMNghZ We’re excited to hear feedback and answer any questions! AMA. https://bit.ly/3TMNghZ September 27, 2024 at 07:57PM

Show HN: A tool that turns everyday computers into your own AI cloud https://bit.ly/3ZDxbit

Show HN: A tool that turns everyday computers into your own AI cloud I have a favour to ask. I’ve been working for a while on Kalavai, a project to make distributed AI easy. There are brilliant tools out there to help AI hobbyists and devs on the software layer (shout out to vLLM and llamacpp amongst many others!) but it’s a jungle out there when it comes to procuring and managing the necessary hardware resources and orchestrating them. This has always led me to compromise on the size of the models I end up using (quantized versions, smaller models) to save cost or to play within the limits of my rig. Today I am happy to share the first public version of our Kalavai client (totally free, forever), a CLI that helps you build an AI cluster from your everyday devices. Our first use case is distributed LLM deployment, and we hope to expand this with the help of the community. Now, the favour! I’d love for people interested in AI at scale (bigger than a single machine) to give it a go and provide honest feedback. Do you share our motivation? If you tried Kalavai, did you find it useful? What would you like it to do for you? What are your painpoints when it comes to using large LLMs? https://bit.ly/3zoGHve September 27, 2024 at 07:12PM

Show HN: Scandium: No-Code Test Automation Tool for Software Testing https://bit.ly/4dq1FaR

Show HN: Scandium: No-Code Test Automation Tool for Software Testing https://bit.ly/4gJAFWz September 24, 2024 at 12:01PM

Show HN NGINIX Manager https://bit.ly/3Buh3Wd

Show HN NGINIX Manager A simple NGINX manager https://bit.ly/4dlJPpy September 26, 2024 at 09:09PM

Show HN: Structured, GitHub App for Automated DBT PR Reviews https://bit.ly/3Y2l1xh

Show HN: Structured, GitHub App for Automated DBT PR Reviews Scaling data teams today means dealing with the complexity of the modern data stack. While DBT has become a core tool for transforming raw data into structured, analytics-ready tables, most teams are using it in ways that lead to chaos: duplicated models, inconsistent metrics, and inefficient SQL that directly impacts cloud spend. The real issue isn’t with DBT itself—it’s in how it’s applied across teams. Here’s the typical setup: Finance defines a revenue model, Marketing calculates customer lifetime value, and Product defines churn. All in DBT, but all with slightly different logic, leading to metric fragmentation. This results in data drift, conflicting reports, and a ton of unnecessary engineering time spent reconciling definitions. Worse, engineers end up re-inventing the wheel by duplicating logic that already exists in other models. The inefficiencies don’t stop there: suboptimal SQL patterns (e.g., full-table scans, poor joins) creep into production and drive up cloud costs. We designed our GitHub App to automate the grunt work of DBT model management, focusing on three key areas: preventing redundant logic, maintaining the semantic layer, and optimizing SQL performance. --- (1) Stop Redundant Models: A lot of teams waste time rebuilding models that already exist. Engineers aren’t aware of what’s been built, so they duplicate work. Our app automatically reviews pull requests, flags redundant models, and suggests reusing existing logic. This keeps your key metrics like revenue or churn consistent across teams and prevents conflicting reports. (2) Maintain the Semantic Layer: DBT’s value is in creating a semantic layer—a consistent definition of business metrics. But as teams scale, maintaining this layer gets tricky. People unknowingly break it with small changes, leading to inconsistencies. Our app checks every new model for deviations from the semantic layer, flagging inconsistencies before they go live. This prevents those all-too-common situations where two departments are debating whose revenue number is right. By ensuring everyone’s using the same definitions, you avoid trust issues with the data. (3) SQL Performance = Real Costs: Bad SQL isn’t just a performance problem—it’s a cost problem. Inefficient joins, full-table scans, and poorly written SQL in your DBT models can blow up your cloud bill. Our app reviews SQL in pull requests, flags inefficiencies, and suggests optimizations. Example: An engineer submits a model that joins two large tables without filtering. Our app flags the full-table scan and suggests using indexed columns and adding WHERE filters. This reduces query cost and improves performance before the code hits production. --- Data engineers are already stretched thin with the demands of modern data pipelines. By automating model consistency checks, semantic layer enforcement, and SQL performance reviews, our GitHub App frees up your team to focus on higher-impact work rather than wasting cycles on repetitive tasks or fighting fires caused by bad data logic. The app is live—give it a try, and let us know how it’s improving your workflow. Also, keep an eye out for our upcoming DBT code generation features—we’re automating more of the heavy lifting soon. https://bit.ly/4eC3fYc September 26, 2024 at 10:28PM

Show HN: Oku – A Web browser with an emphasis on local-first data storage https://bit.ly/4gHOt4a

Show HN: Oku – A Web browser with an emphasis on local-first data storage Hello HN, My name is Emil. I'm a recent unemployed graduate, and I've been spending a lot of my time working on a passion project from my teenage years. When I was younger, I wanted a place on the Web that I could call my own—not a social media page, but a proper site. I was interested in the IndieWeb community for a while, but I grew to believe a P2P alternative to the Web made more sense than having people host federated services. My browser isn't production-ready, but I'm satisfied with the progress so far and would appreciate thoughts & feedback. Thank you! https://bit.ly/3MZQ7QP September 26, 2024 at 02:33AM

Show HN: FastIndex, an open-source search engine indexation tool https://bit.ly/3MXAidB

Show HN: FastIndex, an open-source search engine indexation tool There's a lot of paid search engine indexation tools out there and I wanted to create my own. Been working as an engineer for over a decade now and my open-source contribution has always been something I wanted to do. Thus I decided to create FastIndex, an open-source search engine indexation alternative to paid solutions such as TagParrot, URLMonitor, Omega Indexer and many more. Source: https://bit.ly/3MWJhf3 Wiki: https://bit.ly/3N1Pd6q https://bit.ly/3MWJhf3 September 26, 2024 at 02:16AM

Show HN: A new, improved, and open-source clipboard history Chrome extension https://bit.ly/3XWudnI

Show HN: A new, improved, and open-source clipboard history Chrome extension Hey all! I recently built a new, improved, and open-source chrome extension that allows you to access, track, and manage your clipboard's history! It's very useful for things like refactoring code, finding obscure commands, and backing up form data. Existing solutions are either closed-source, slow, riddled with ads, or all of the above. My goal is to make this extension the most trustworthy, performant, and easy to use version! Your feedback would be very much appreciated, thanks! Extension: https://bit.ly/3XYvAm7... Repo: https://bit.ly/3XVBLay https://bit.ly/3zBb9lN September 26, 2024 at 12:51AM

Show HN: Public Domain Torrent Site https://bit.ly/3MYyTmU

Show HN: Public Domain Torrent Site I have been working on this site for 10 years. It is a BitTorrent Indexer that uses WebTorrent to make public domain educational media freely available on the Internet. https://bit.ly/3XYtHpA September 25, 2024 at 10:37PM

Show HN: A toxic conversation with ChatGPT for research https://bit.ly/3THCn0T

Show HN: A toxic conversation with ChatGPT for research I started this conversation with ChatGPT as an exercise to explore the ways in which young folks can be influenced by peers into toxic ways that may lead them astray. So instead of asking generally, I went for a devil's advocate approach. I am actually amazed and impressed by how suggestive and understanding of current youthful traits that ChatGPT is aware of. Any suggestions for other questions? When you pose a question, let me know what you intend on establishing from the response. https://bit.ly/3TJShrI September 25, 2024 at 05:08AM

Show HN: Chrome extension to summarize HN comments using AI and LLMs https://bit.ly/3BjkdfG

Show HN: Chrome extension to summarize HN comments using AI and LLMs Hello! I built this to solve a personal problem of where I didn't want to wade down large chains of HackerNews comment threads in order to get the key takeaways from the discussions. I built this Chrome extension, which supports both OpenAI and Ollama(local LLMs) to summarize comments and display the summary within Chrome's sidepanel. It's open source as you can see from the link, feedback appreciated! https://bit.ly/4ezQzRI September 25, 2024 at 02:16AM

Show HN: Broken Hill: A Productionized GCG Attack Tool for Use Against LLMs https://bit.ly/4dnGwxU

Show HN: Broken Hill: A Productionized GCG Attack Tool for Use Against LLMs https://bit.ly/4ddDgFm September 24, 2024 at 11:31PM

Show HN: Oodle – serverless, fully-managed, drop-in replacement for Prometheus https://bit.ly/4efFC7R

Show HN: Oodle – serverless, fully-managed, drop-in replacement for Prometheus Hello HN! My co-founder, Vijay and I are excited to open up Oodle for everyone. We used to be observability geeks at Rubrik. Our metrics bills grew like 20x over 4 years. We tried to control spend by getting better visibility, blocking high cardinality labels like pod_id, cluster_id, and customer_id. But that made debugging issues complicated. App engineers hated blocking metrics, and blocking others' code reviews was not fun for platform engineers either! Migrations (and lock-ins) were very painful, the first migration from Influx to Signalfx took 6+ months and the second migration from Splunk took over a year. Oodle is taking a new approach to building a cost-efficient serverless metrics observability platform. It delivers fast performance at high scale. We leverage custom storage format on S3, tuned for metrics data. Queries are serverless. The hard part is how to achieve fast performance while optimizing for costs (every cpu cycle, storage/memory byte counts!). We've written about the architecture in more detail on our blog: https://bit.ly/4gA0RTC... Try out our playground with 13M+ active time series/hr & 13B+ samples/day: https://bit.ly/3BhOYBF Explore all features with live data via Quick Signup: https://bit.ly/47EawEA - Instant exploration (<5min): Run one command to stream synthetic metrics to your account - Easy integration (<15min): Explore with your data from existing Prometheus or OTel setup. We’d love your feedback! cheers https://blog.oodle.ai/building-a-high-performance-low-cost-metrics-observability-system/ September 24, 2024 at 01:39PM

Show HN: OpenFreeMap – Open-Source Map Hosting https://bit.ly/3BhOHyD

Show HN: OpenFreeMap – Open-Source Map Hosting Hi HN, After 9 years of running my own OpenStreetMap tile server infra for MapHub ( https://bit.ly/3XypM17 ), I've open-sourced it and launched OpenFreeMap. You can either self-host or use our public instance. Everything is open-source, including the full production setup — there’s no 'open-core' model here. Check out the repo ( https://bit.ly/3zBqYc1 ). The map data comes from OpenStreetMap. I also provide weekly full planet downloads both in Btrfs and MBTiles formats. I aim to cover the running costs of the public instance through donations. Looking forwards for your feedback. https://bit.ly/4e8VCbW September 24, 2024 at 12:59PM

Show HN: An expression parser supporting multiple types https://bit.ly/3TENoQA

Show HN: An expression parser supporting multiple types This C library is part of a main project aimed at providing a reactive key-value (KV) database. The data is typed (numbers, strings, dates, or booleans) and can include formulas with references to other entries. Clients connected to this database receive a real-time data stream with updates to the subscribed keys, allowing them to react to changes and their dependencies. Essentially, it’s like building a distributed Excel, where data and formulas dynamically update across the system. I couldn’t find any libraries that offered the full set of features I needed for evaluating expressions, so I decided to create my own. This sub-project is open-source and available on GitHub. Feedback is welcome! https://bit.ly/3XTTD5p September 24, 2024 at 11:30AM