ads
Saturday, October 31, 2026
Show HN: Ghostman – A fast, keyboard-driven TUI API client and webhook inspector https://ift.tt/s7okLAJ
Show HN: Ghostman – A fast, keyboard-driven TUI API client and webhook inspector https://ift.tt/q5lpb0A October 30, 2026 at 02:48PM
Friday, October 30, 2026
Show HN: I rebuilt Kubernetes for ESP32 microcontrollers https://ift.tt/GZvb5g2
Show HN: I rebuilt Kubernetes for ESP32 microcontrollers https://ift.tt/5sfUQqM October 29, 2026 at 11:35PM
Thursday, October 29, 2026
Show HN: Radix - A WebGPU-powered 3D Gaussian Splat renderer built on Three.js https://ift.tt/mN67FBH
Show HN: Radix - A WebGPU-powered 3D Gaussian Splat renderer built on Three.js https://ift.tt/gYrmfPV October 30, 2026 at 12:10AM
Show HN: Witnessory – a signed record of the web's move to post-quantum crypto https://ift.tt/gMilQA5
Show HN: Witnessory – a signed record of the web's move to post-quantum crypto https://witnessory.org October 29, 2026 at 11:40PM
Wednesday, October 28, 2026
Show HN: SlideBySide – Present slides without sharing your screen https://ift.tt/xEdFMpw
Show HN: SlideBySide – Present slides without sharing your screen Why is it useful?
- Present slides without sharing the screen and have every one be on the same page.
- Present together with other cohosts and all control the current page.
- Other features such as presentation collab features. Why should I not just vibe code my own solution?
- Its a hassle to maintain a version online. https://ift.tt/taA1Gho October 29, 2026 at 12:07AM
Show HN: Generate (animated) QR Codes straight in your terminal https://ift.tt/CQMnXN4
Show HN: Generate (animated) QR Codes straight in your terminal https://ift.tt/fHBUJia October 28, 2026 at 11:26PM
Tuesday, October 27, 2026
Show HN: Gifup – GIF tools that run in the browser (ffmpeg.wasm) https://ift.tt/iyfO4u7
Show HN: Gifup – GIF tools that run in the browser (ffmpeg.wasm) Hi HN. GIFUP was an avatar and glitter generator I ran from 2007; it made about 1.3M animations before the web moved on. This year I rebuilt the domain as a set of GIF tools where the encoding happens in the tab via ffmpeg.wasm: video to GIF, compress, resize, crop, captions, GIF to MP4, plus presets that fit Discord, Slack, Twitch and WhatsApp limits. Nothing is uploaded. The file is read from disk, FFmpeg runs in WebAssembly, and the result is handed back as a blob. The engine is about 10 MB compressed, fetched once and cached. The core is single-threaded because the threaded build needs COOP/COEP headers, so a 10-second 720p clip takes a few seconds rather than a fraction of one. No size caps, no queue, no account. Two optional server features: "Get a link" hosts a finished file if you want to paste it in a forum, and a WhatsApp sticker page can deliver a sticker to your phone. Both are off until you press the button. The glue code on the page is plain JS; the FFmpeg build is the unmodified @ffmpeg/core 0.12.6 (GPL, libx264 and libx265 included). I would like to hear where it breaks, especially on Safari, and which tool you are missing. https://gifup.com/ October 27, 2026 at 10:40PM
Monday, October 26, 2026
Show HN: Hob - Open-source, on device alternative to Claude Cowork and ChatGPT https://ift.tt/IcXhjQk
Show HN: Hob - Open-source, on device alternative to Claude Cowork and ChatGPT Hey HN, I'm Christian and this is a new OSS project I've been noodling on for the last couple of weeks. Hob is a free, MIT-licensed desktop app for Mac, Windows and Linux that works on your files the way the Claude and ChatGPT desktop apps do, except the model runs on your own computer. It uses open-weight models (currently Gemma 4 and gpt-oss). All your chats and files never leave your system. The only things that go online are web search (can be turned off in settings) and any MCP connectors you add yourself. Main features: - Automatically finds the best open-weight model for your hardware on first install (Mac/Linux/Windows) - Writes sandboxed Python code for ad hoc tools (data processing, document conversion, graphs etc.) - Built-in web search (powered by SearXNG) - MCP support - Image recognition - CRUD to your local filesystem - Works fully offline with no internet Before building this I tried every local AI app I could find, but it seemed each one had different problems. The main thing I noticed is that none of them could do what the ChatGPT and Claude apps do with your files, e.g. read and change files in a folder, look at images, run code in a sandbox, draw a chart. The most advanced thing any of them did was paste a plain text file into a small context window, which basically failed as soon as you had a file larger than a few KB. I couldn't find anything that could easily parse images and bigger PDFs. The OSS solutions that were more advanced all seemed to default to hosted models (either OSS or closed), which defeats the entire point of the privacy that on-device OSS offers. They also seemed to push towards an account/open-core style business. I build it using Electron bridged with llama.cpp in its own process. On initial load, the system will try and figure out if you have a compatible graphics card and if not default to CPU-only (I ran into a lot of issues of graphics card beeing seemingly compatible but then crashing on first query, there's a whole saga to this lol). Code execution happens in a sandbox that can only write to the chat's folder and has no network unless asked (sandbox-exec on macOS, Landlock on Linux and a low-integrity process on Windows). Images go to a small Qwen3-VL model. Dictation is a local Whisper model. The hardest part was definitely to get the smaller models to use tools reliably. Ran into a ton of issues of 4B models happily saying "here's your chart" without ever calling the chart tool or get stuck in an infinite call loop. I spent a fair amount of the work building guards vs tweaking prompts (that didn't really work for small models). I understand the pushback that local models are weaker than frontier cloud models. This is true, but for most tasks for most normal people, I'd argue you rarely need the newest frontier model. From my own evals, systems with 16+ GB of RAM do really well, and if you have a GPU or more RAM, you get very close to what you'll get out of the box with a frontier hosted model. Excited to hear what you think and see how it runs on other systems! https://ift.tt/6ZpxmdP October 26, 2026 at 11:12PM
Show HN: RPI Pico (RP2040) PIO Emulator Package for Go https://ift.tt/nWKEgFp
Show HN: RPI Pico (RP2040) PIO Emulator Package for Go I have been developing a bus communication protocol for a while using the Raspberry Pi Pico (RP2040)'s programmable IO (PIO) hardware (basically- 8 tiny processors that run assembly programs to manipulate GPIO pins) and I needed some way of testing my programs, as the real hardware is undebuggable. Unfortunately, none of the existing emulators I could find supported connecting multiple RP2040s together, and they each had a few different missing bits of functionality, so I decided to make my own! My emulator supports running multiple RP2040s electrically connected together, and emulates the whole of their PIO blocks (all 8 processors each), along with all of the more niche inter-processor interactions and GPIO functionality I could possibly find in the datasheet. I built it as a Go package so that it can be used inside of other programs; I intend on using it myself inside a test suite against my in-development protocol. I've spent quite a long time checking and double-checking most of the logic so it is as complete and accurate as I can make it, but if you spot anything amiss, I'd really appreciate the feedback. Thanks for checking it out!
Hope it can be useful to other RP2040 users. https://ift.tt/5Mgs89T October 27, 2026 at 12:29AM
Show HN: HealthViewer – A local-first desktop app for tracking lab results https://ift.tt/V3ApQEi
Show HN: HealthViewer – A local-first desktop app for tracking lab results https://ift.tt/cEw7K0e October 26, 2026 at 08:56PM
Sunday, October 25, 2026
Show HN: Hurston-1 open source TTS that speak Southern American and Afro-English https://ift.tt/RlUu8KJ
Show HN: Hurston-1 open source TTS that speak Southern American and Afro-English Hey HN, I’m open sourcing my Hurston-1 narrowband TTS model (Apache-2.0). It is very good with American English and does especially well with Southern American and Regional African English accents. Destined AI is based in Atlanta, and I built this because I’m amazed by what AI can do, and I wanted to help introduce and welcome more people to feel comfortable using and interacting with AI. Since 2021, I’ve been collecting fully consented voice data from over 20,000 people, 1M audio recordings, and over 1,800 hrs of audio captured to date. For this model, I used a small part of the data, 136 hrs, to finetune Maya1 and expand it’s capabilities. Hurston-1 is a narrowband TTS model that focuses the output on telephone use because that is still where most of these voices will be heard. I’ve used it to support customers such as Amazon and Salesforce and generated over half-mill revenue with it. Now I’m open sourcing it to help unlock opportunities for others as well. Limitations: this model doesn’t maintain the exact same voice across prompts and doesn’t support voice cloning. For many use cases, this is fine but maybe not so - if you’re creating podcasts. This is the first release - I’m already working on the next version with improvements. I would appreciate, if anyone is interested in this space, download the model from huggingface, check it out and let me know any feedback. Also, open to connecting with others in this space as well. https://ift.tt/9wegXTx October 25, 2026 at 08:30AM
Saturday, October 24, 2026
Show HN: Duls – A bit of du, a bit of ls. Best of both https://ift.tt/U2eCIxh
Show HN: Duls – A bit of du, a bit of ls. Best of both https://ift.tt/Ddw0s1G October 25, 2026 at 03:14AM
Show HN: A phone lock I can't undo on impulse because I never learn the code https://ift.tt/Zke9BVH
Show HN: A phone lock I can't undo on impulse because I never learn the code https://ift.tt/x64tFDJ October 24, 2026 at 09:22PM
Show HN: Agentsh – A Decentralized, Scalable Multiagent Harness https://ift.tt/oD3BEm5
Show HN: Agentsh – A Decentralized, Scalable Multiagent Harness https://ift.tt/sd6KxP8 October 25, 2026 at 01:19AM
Friday, October 23, 2026
Show HN: Swurf – Swipe through the real web https://ift.tt/vwEklb4
Show HN: Swurf – Swipe through the real web Hello HN, Swurf allows you to swipe through the real web one page after another. It learns what you like and doesn’t show you the same page twice. The idea came naturally: there’s a seemingly infinite amount of high quality blog posts, random topic rabbit holes, and trending or cool sites that exist which the average person simply will never load on their browser. Swurf solves that by allowing you to rapidly traverse a constantly growing and learning feed of sites that it thinks you’d want to view. Quickly sift through endless pages and find things you like - effortlessly. One of the core drivers for making this project was the desire for real content. The project started as a simple tool for myself to find new and interesting things to learn and read about. Hours spent scrolling on my phone - while it still has its place - is increasingly losing its sense of reward. It’s genuinely hard to put into words exactly how swurfin’ feels but it’s almost refreshing in a sense. I encourage you to try it out and provide feedback. At this time there are no accounts and you can jump in and start swurfin’ immediately. There is a three dot menu button on the bottom right which allows you to submit links, save links, control adult content, and search similar topics. One of the things that’s great about Swurf is that its almost like a group browsing session (though entirely private) where everyone else who is spending time hunting down great reads or solid sites is helping massage the feed for you and you for them - all within your own set of interests. Enjoy and please ask any questions you have as theres plenty of room to grow. https://swurf.org/ October 23, 2026 at 01:54PM
Show HN: Vericto – Check AI agents' SQL before Postgres or MySQL runs it https://ift.tt/zD4Kd2t
Show HN: Vericto – Check AI agents' SQL before Postgres or MySQL runs it https://ift.tt/t2vMi6Q October 23, 2026 at 09:16PM
Thursday, October 22, 2026
Show HN: Fire hazard zone, power line risk and fire history for any CA address https://ift.tt/5hRbAUX
Show HN: Fire hazard zone, power line risk and fire history for any CA address https://ift.tt/t4IMnKN October 22, 2026 at 10:52PM
Show HN: OnKey – I'm a bad singer, so I built an app to show where I'm off-key https://ift.tt/d5x8sFE
Show HN: OnKey – I'm a bad singer, so I built an app to show where I'm off-key I'm not a good singer, but strangely enough I still like singing. I've improved a little over time with singing apps, but their song lists are limited by licensing, so I'd often open a song expecting the original and find a cover I didn't want to sing to. So I built OnKey: a free, open-source (MIT) Mac app that listens to whatever you're playing in Spotify, Apple Music or YouTube, learns the singer's melody, and shows you live where you're sharp or flat, with a score at the end. Nightingale did something similar years ago for your own audio files. OnKey works with the music you already stream. Demo (1:35, sound on): https://singonkey.app The first time you play a song, you can sing along while it listens, or play a whole playlist and let OnKey learn the songs in the background before you come back to sing them. When a song finishes it learns the notes on your Mac (a couple of minutes), and from the second play the notes scroll towards you before you sing them. Learning the song and scoring you both happen on your Mac. No account, no analytics, and the music is never saved or sent anywhere. Your takes stay there too. To check the score means something, I ran 20 recordings of singers, from poor to professional, that music experts had already scored. OnKey ranked them in nearly the same order as the experts did. It's still rough in places: busy mixes and harmonies can confuse which voice is the tune, and now and then a note lands an octave off. Scoring is forgiving on purpose for now, set for beginners like me, so stricter modes for better singers are next. And it's Apple Silicon and macOS 14.2+ only. I'd love to hear where it scores you unfairly or the notes go wrong, and I'm happy to answer questions about how it hears the music and learns a song. https://ift.tt/CP3YbDK October 22, 2026 at 09:04PM
Wednesday, October 21, 2026
Show HN: Gembop, a block-placement puzzle game in 2D and 3D https://ift.tt/erBcEhM
Show HN: Gembop, a block-placement puzzle game in 2D and 3D https://gembop.games/play October 21, 2026 at 10:04PM
Show HN: ULDA – Client-side encryption with a simple JavaScript API https://ift.tt/DXrnjql
Show HN: ULDA – Client-side encryption with a simple JavaScript API Hi there! We have designed client side encrypted storage that sync on multiple devices and allows to hold that data on your server. We have started as chat application in late 2022 that has completly open-readable database. The idea was simple - server holds zero secrets, like non at all. But we ran into the problem of authentication - there were no auth method what could make quick, synced authorization with completely zero secrets. So in early 2023 we came to an idea of authentication via proof of ownership by providing proof that next step is made by the same user. Now you can see our job as ulda core https://ift.tt/KNeFJsH . Ulda was called as universal liniar data authentication. But a year later we faced that “None need another extremely secure chat application” and we decided to export our ideas and share ulda as a method in order to gather some feedback from some smarter people. In 2025 we tried to publish our white papers at REKS, CMIS and in the end in Journal of Cryptology. Ulda was never published and then we decided that it would probable a good idea to go and publish at least something somewhere to bring some attraction and hopefully some feedback. So, we assembled an sdk, that worked via ulda core and gave developers an ability to work in with encrypted and synced storage directly by editing json. We tried to make is simple as possible. We hope to bring some attention not to a sdk itself, but rather to authentication method itself. You can read our white paper here: https://ift.tt/NfdVuZa Also, if you would like to see our sdk, you can also try it here: https://docs.ulda.io , https://ulda.io . Thank you very much for your time and we hope for your feedback about authentication method. Best regards,
Mark Shaposhnik
Vitalii Radko https://demo.ulda.io/ October 20, 2026 at 12:16AM
Tuesday, October 20, 2026
Show HN: Skillsync – Coding agent memory but it's just search https://ift.tt/zO9Efsd
Show HN: Skillsync – Coding agent memory but it's just search I built a claude mod that searches every session (across harnesses) on each prompt. On every prompt, it runs a series of decision models to pick the right sessions + messages and adds anything relevant to the context. Requires no indexing or setup and works out of the box Demo: https://youtu.be/o5vEf8tgus4 https://skillsync.com October 21, 2026 at 12:04AM
Show HN: Agent Archive – Session Handoffs and Backups for Claude Code and Codex https://ift.tt/DFRfCZ0
Show HN: Agent Archive – Session Handoffs and Backups for Claude Code and Codex I often switch between Claude Code and Codex throughout the week, and between my laptop and an older machine I leave running at home. I built agent-archive to make it easier to pick up where I left off, especially when i hit usage limits. It lets you (or a coding agent) find past coding sessions and hand their conversation context to another agent. You can try local handoffs without setup, or connect your own S3/R2 bucket for automatic archiving and access from other machines. It also tracks usage across agents and gives you a full archive of your session transcripts across agents, which I've found useful to iterating and improving my AGENTS.md and lint rules. https://ift.tt/xIyuDJW October 20, 2026 at 10:54PM
Monday, October 19, 2026
Show HN: RawLink – Building a VPN protocol from scratch https://ift.tt/HUEdBin
Show HN: RawLink – Building a VPN protocol from scratch https://ift.tt/9lxEb0N... https://ift.tt/D4z7eZk October 20, 2026 at 01:59AM
Show HN: ECG Signal Processing Starter Kit – Python, SciPy, and FastAPI https://ift.tt/gOBQnSN
Show HN: ECG Signal Processing Starter Kit – Python, SciPy, and FastAPI https://ift.tt/c9eELJd October 19, 2026 at 09:30PM
Sunday, October 18, 2026
Show HN: Kramena – Classic books, 20 minutes a day, with spoiler-safe recaps https://ift.tt/HmBPsgA
Show HN: Kramena – Classic books, 20 minutes a day, with spoiler-safe recaps https://ift.tt/ga1rvZn October 18, 2026 at 11:02PM
Saturday, October 17, 2026
Show HN: Parin – A 2D game engine between raylib and Godot, written in Dlang https://ift.tt/KZ6w5FU
Show HN: Parin – A 2D game engine between raylib and Godot, written in Dlang https://ift.tt/vIOyRm4 October 16, 2026 at 08:22PM
Show HN: Flying rocket car x Flappy Bird https://ift.tt/gn0MzXV
Show HN: Flying rocket car x Flappy Bird https://ift.tt/6yj5DxW October 18, 2026 at 12:14AM
Friday, October 16, 2026
Show HN: MacVisor Runs VMs with VLANs and Firewall on Apple Silicon https://ift.tt/b1Z9ez2
Show HN: MacVisor Runs VMs with VLANs and Firewall on Apple Silicon https://ift.tt/G3Xctdu October 16, 2026 at 09:46PM
Thursday, October 15, 2026
Show HN: Council AI. Spot biases / hallucinations in top AI models https://ift.tt/Wdwuhen
Show HN: Council AI. Spot biases / hallucinations in top AI models Examples:
- The best and worst of AI in 2026. https://ift.tt/Wj8AISo
- The best and worst of Trump 2026. https://ift.tt/Ku5bDrl https://councilai.app October 15, 2026 at 11:19PM
Wednesday, October 14, 2026
Show HN: I wrote an email app that accepts only one sender per inbox https://ift.tt/m1nLdrz
Show HN: I wrote an email app that accepts only one sender per inbox Infinite inboxes, including prefixes where you don't have to create inboxes before you use them -- they self create when you add a suffix. Three types of inboxes:
- One sender. Locks to the first domain to send email (except gmail, outlook, etc where it locks to an email account for obivous reasons. It bounces everyone else.
- Temporary. Lasts for a while (24h, 7days, etc) then bounces. Received messages stay or go as per your choice.
- Prefix. The most powerful. Declare a prefix like "service" and then invent any suffix, like "service-netflix" or "service$bank" and it creates a new one-sender inbox locked to whoever wrote first. Bring you own domain. Bring your own AI (MCP server ready). Forwarding. Please help me test it. https://clova.email/ October 15, 2026 at 12:00AM
Show HN: Mixdog – stretch your Claude/ChatGPT subscription with 63% fewer tokens https://ift.tt/wT9IVog
Show HN: Mixdog – stretch your Claude/ChatGPT subscription with 63% fewer tokens https://ift.tt/oDdgZUl October 14, 2026 at 11:33PM
Tuesday, October 13, 2026
Show HN: XY Music Notation https://ift.tt/Tgh6KID
Show HN: XY Music Notation I've been playing around with a new kind of isomorphic music keyboard that uses a base-12 notation, and wanted some sheet music to learn some new music for it. The idea is to represent each note in 12-TET as a number, so N ∈ (0,1,2,3,4,5,6,7,8,9,X,Y). I used codex to build a music score rendering backend based on MuseScore with custom note heads that display the pitch class of each note. I've found it pretty helpful for learning how to play my own instrument, so wanted to share in case anyone else was interested. https://music.xy.dev/ October 14, 2026 at 12:43AM
Show HN: How Random Are You? https://ift.tt/vkjwpBV
Show HN: How Random Are You? https://ift.tt/qt8oN0c October 13, 2026 at 11:57PM
Show HN: Knitting – A shared-memory concurrency runtime for Node.js,Bun,and Deno https://ift.tt/NGf7A3U
Show HN: Knitting – A shared-memory concurrency runtime for Node.js,Bun,and Deno https://ift.tt/CDZxSBN October 13, 2026 at 10:19PM
Monday, October 12, 2026
Show HN: Closed Form – OpenAI's 372 math results, explained in plain English https://ift.tt/Xy7P9iw
Show HN: Closed Form – OpenAI's 372 math results, explained in plain English https://ift.tt/c7kH5Pl October 12, 2026 at 07:30PM
Show HN: IHMT – Infinite Hierarchical Memory Tree for AI Agents https://ift.tt/envUJcb
Show HN: IHMT – Infinite Hierarchical Memory Tree for AI Agents https://ift.tt/6zsNWKb October 12, 2026 at 11:34PM
Sunday, October 11, 2026
Show HN: THX-01 – Open-source decision model, free API, 60M+ tokens processed https://ift.tt/h3yQJVG
Show HN: THX-01 – Open-source decision model, free API, 60M+ tokens processed https://twitter.com/reversedoofz/status/2109228322752471304 October 12, 2026 at 04:31AM
Show HN: Kindlefall a 3D ARPG in the browser, built with Three.js https://ift.tt/FcBpSfL
Show HN: Kindlefall a 3D ARPG in the browser, built with Three.js https://ift.tt/b58KWYc October 12, 2026 at 01:38AM
Show HN: HarmoNova: A modern, customizable Hacker News reader for Chrome https://ift.tt/h7nuevV
Show HN: HarmoNova: A modern, customizable Hacker News reader for Chrome https://ift.tt/8DMPl0d October 12, 2026 at 01:42AM
Show HN: OpenClear – Self-hosted check and ACH fraud screening for small banks https://ift.tt/hbrfiIN
Show HN: OpenClear – Self-hosted check and ACH fraud screening for small banks https://ift.tt/uDb5VYF October 11, 2026 at 11:14PM
Saturday, October 10, 2026
Show HN: Throw Alice and Bob into Sagittarius A* https://ift.tt/8gCGPxr
Show HN: Throw Alice and Bob into Sagittarius A* Black Hole Lab is an open source physics simulation of a rotating black hole and two observers, Alice and Bob. You can launch Alice and Bob on paths around or into the black hole and watch events play out. The simulation lets you see a slice of the physics with one or two dimensions of space plus one of time. As a fun extra, Alice and Bob each have a beacon that can transmit a radio wave in all directions. The simulation calculates and shows the wavefront of these signals, accounting for the warping of space and time by the black hole. The simulation also records whenever Alice and Bob receive each other's transmission. The simulation takes pains to respect Einstein's equations of general relativity. To the precision of the numerical integration, the effects shown in Black Hole Lab are really what the math of general relativity says will happen. Black Hole Lab specifically uses the Kerr metric, which is an exact solution to
Einstein's equations of general relativity for a rotating mass. That said, physicists agree that general relativity cannot be the final word on black hole physics. Nobody can say what would really happen to Alice and Bob. My original goal for black hole lab was to try to understand the strange time compression effects at the inner horizon, which physicists call infinite blueshift. IMHO, Black Hole Lab succeeds in showing this effect in an understandable way. I added a specific discussion of infinite blueshift in the readme. https://ift.tt/L1VdtNm October 11, 2026 at 04:28AM
Show HN: A functioning Celery Man computer https://ift.tt/a0Jls46
Show HN: A functioning Celery Man computer Hi HN! It dawned on me that the latest AI models (Minimax H3 Max / Audio, GPT-6 Luna) can finally recreate the computer from the classic Tim & Eric Celery Man sketch. I spent a few days creating it: it has 100% parity with the original script, plus anything you ask for (doggy woof woof man or potato boy). Happy to answer any questions about the pipeline! https://celeryman.fun/ October 11, 2026 at 01:08AM
Show HN: An agent harness for long-running pentests and code audits https://ift.tt/VpaAz4t
Show HN: An agent harness for long-running pentests and code audits https://ift.tt/xNu7rB5 October 10, 2026 at 10:59PM
Friday, October 9, 2026
Show HN: Let your AI agents paint big arrows, boxes and text on your screen https://ift.tt/Gwj6Pbg
Show HN: Let your AI agents paint big arrows, boxes and text on your screen https://ift.tt/SBvYbOV October 9, 2026 at 06:03PM
Show HN: Coolphabets – Teaching kids to read and write in their mother tongue https://ift.tt/tremJUW
Show HN: Coolphabets – Teaching kids to read and write in their mother tongue COOL + ALPHABETS = COOLPHABETS COOLPHABETS teaches you to read and write in your mother tongue - for kids who are just starting, and adults who never did. Millions of families live outside the country they were born in. While parents can read and write in their native language, their kids - the second generation - can barely do so. Kids often speak their mother tongue fluently because it is the language they use at home. But reading and writing is rarely a "felt need" for them. And kids are unlikely to feel that need unless parents nudge them - preferably at an early age. Many parents genuinely want their kids to be able to read and write in their mother tongue. This usually means a textbook or a weekend language class. Both can feel like a chore. Most language apps teach you to 'speak' a language. COOLPHABETS teaches you to 'read' and 'write'. Reading and writing are a genuinely difficult problem to solve - especially on an app. Getting a child to write letters correctly is the harder part. That's where we built CoolphaSense - a collection of tiny AI models that run entirely on the device and provide instant feedback on handwriting. COOLPHABETS is designed for kids. No ads. No tracking. COPPA compliant. The apps are designed to work fully offline. No cloud roundtrips to a 'Mothership AI model'. COOLPHABETS is currently available in Indian languages. - Telugu
- Kannada
- Hindi
- Tamil
- Marathi Most other Indian languages will be onboarded by December 2026. Other world languages (Asian, Middle Eastern, European) are early next year. The bigger motivation is cultural. A language can survive for several generations as a spoken language, but if its script is lost, something important is lost with it. If one generation stops reading and writing in its native script, reconnecting with that part of the culture becomes much harder. If you are a parent who feels this way, COOLPHABETS is for you and your kids. https://ift.tt/OX0q3pG Android: https://ift.tt/t91oB7F... iPhone: https://ift.tt/Ml60bDJ We would love to hear from parents settled outside their native countries, parents raising kids in multilingual households, and from people working on handwriting recognition, on-device AI/ML, and language learning. October 9, 2026 at 11:05PM
Thursday, October 8, 2026
Show HN: Pocketty – iPhone SSH terminal that pings you when an agent is blocked https://ift.tt/ErMOfZe
Show HN: Pocketty – iPhone SSH terminal that pings you when an agent is blocked Hello~! pocketty is an SSH terminal for iPhone and iPad, made for herdr. herdr keeps your agent panes alive on your computer and knows the state of each one: working, needs you, or done. I made this in anger/desperation for the latter half of my recent paternity leave. Nap traps are sweet, but there's only so much doom-scrolling and movie-watching I can handle... In any event, I've been using it for the last couple months and no longer have to be my desk anymore to be productive. Now the nap-traps are still productive (when i want them to be) ! How it works: - The app talks to your computer directly over plain SSH. Tailscale is the easy way to reach it from anywhere but any SSH host you can reach works. - A small Rust daemon on the host watches herdr. When a pane needs you, it seals the alert to your phone's key with HPKE (X25519, ChaCha20-Poly1305). A stateless relay (pocketty's) on Cloudflare Workers passes the sealed bytes to APNs (Apple Push Notification servers), and a notification extension opens them on the phone. The relay can't read them and keeps nothing. - Your SSH key is made in the Secure Enclave and can't be exported. - The terminal uses libghostty-vt for state and draws with wgpu on Metal, so full-screen TUIs look like they do on your desk, albeit narrower. - Diffs for each agent turn, a file browser, and previews of `localhost` dev servers your agent starts, all through the same SSH connection. No port forwarding to set up. herdr and the daemon are optional. Without them, it's a normal SSH client. There's no account to set up and no analytics or tracking in the app.
The app runs a 14-day free trial, with full feature access. Then, if you're as happy as I am with it, then it can be yours forever with a one-time purchase: $99 for the first two weeks (launch promo) then $129 after that. Happy to answer anything about the sealed push setup or running libghostty on iOS. https://pocketty.app/ October 9, 2026 at 01:11AM
Show HN: K10s – A Clickable Kubernetes TUI (Go, Bubble Tea) https://ift.tt/R3CDsUB
Show HN: K10s – A Clickable Kubernetes TUI (Go, Bubble Tea) https://ift.tt/CpIjw5M October 9, 2026 at 01:29AM
Show HN: Porch – a Mac menu bar work log for Claude Code and Codex sessions https://ift.tt/Ouhvoql
Show HN: Porch – a Mac menu bar work log for Claude Code and Codex sessions https://ift.tt/ciUmGKf October 8, 2026 at 08:57PM
Show HN: AI SRE Arena, an Open Benchmark for AI SRE Agents on Kubernetes https://ift.tt/BhK1oi5
Show HN: AI SRE Arena, an Open Benchmark for AI SRE Agents on Kubernetes https://ift.tt/e4TYzvn October 9, 2026 at 12:13AM
Wednesday, October 7, 2026
Show HN: Durable Actors – OSS Durable Objects with configurable compute https://ift.tt/BuKpvCD
Show HN: Durable Actors – OSS Durable Objects with configurable compute Hi HN, we're Thomas and Olivier from Terse ( https://ift.tt/Zu0DlU1 ) We've built Durable Actors, an open-source alternative to Cloudflare's Durable Objects. A Durable Object/Actor is a tiny server that handles one request at a time and has its own SQLite database. There's exactly one of each in the world and it is addressed by name. This is the perfect primitive for deploying multiplayer agents. Each agent can have its own Durable Actor, and each user can connect to that Actor via websocket. This is fully horizontally scalable. Your users can deploy and share agents at will without putting pressure on a central DB or websocket server. Durable Actors are also great for coordinating agents within a system. Since only one request is handled at a time, you can protect critical data such as a CRM and allow multiple agents to run concurrently without worrying about data races. The only alternative to this is Cloudflare's Durable Objects. However, there is extreme lock in (they pull you into D1, R2 + workers as well) and it wasn't originally built for agentic workfloads when it was released 5 years ago. Some notable projects built on Durable Objects include RampInspect, OpenInspect as well as the multiplayer frameworks Liveblocks and PartyKit. You can now build these kinds of projects on Durable Actors. Durable Actors is a version of DO that is built for concurrent agentic workloads. It is fully open source (MIT License) and includes a helm chart for you to easily self-host. Some key features: - Configurable compute: Specify CPU, RAM, data residency, idle-timeouts all in a decorator - No outer worker: We generate a type-safe client that you can just plug into your existing tech stack. - (coming soon, like today) Export SQLite table via CLI + MCP for exposing OLTP logs to your agent to help you debug. And our Performance Numbers (all p95): - Durable write: 85.6ms - Stateful Read (data in sqlite): 2.14ms - Actor Warm up: 334ms Here is a little counter demo so you can see the latency yourself: https://ift.tt/fDhOSqr Would love your feedback! https://ift.tt/7PXQnEa October 6, 2026 at 10:58PM
Tuesday, October 6, 2026
Show HN: Session-continuity, cryptographic proof of session chains https://ift.tt/JBlCiRK
Show HN: Session-continuity, cryptographic proof of session chains https://ift.tt/vReUEfS October 6, 2026 at 09:52PM
Show HN: AI Circuit Breaker – Reverse proxy to stop agent infinite loops https://ift.tt/3ZaHLPr
Show HN: AI Circuit Breaker – Reverse proxy to stop agent infinite loops https://ift.tt/2VqcOpB October 6, 2026 at 10:59PM
Monday, October 5, 2026
Show HN: Nightwatch – a Mac menu-bar app that tells you when tonight is clear https://ift.tt/afEegoj
Show HN: Nightwatch – a Mac menu-bar app that tells you when tonight is clear https://ift.tt/m2p0lWF October 4, 2026 at 04:31PM
Show HN: I finetuned 1.5B Qwen to near GPT-4o level bash generation perf https://ift.tt/DGAwtOk
Show HN: I finetuned 1.5B Qwen to near GPT-4o level bash generation perf Purely a hobby side project to see how far I can push a really small model, using (mostly) automated training pipelines Original mention: https://ift.tt/uIQYNrE There were a bunch of requests to release it. Full synthetic data: https://ift.tt/PhZCXu4 Models: https://ift.tt/3FgiTKq and https://ift.tt/1DKyrxt Cli https://ift.tt/YoCKymf feel free to train/use the data as you wish. https://ift.tt/weqRDBm October 5, 2026 at 10:35PM
Sunday, October 4, 2026
Show HN: Jev-pages – Building landing pages real time with Jev https://ift.tt/y2cTopi
Show HN: Jev-pages – Building landing pages real time with Jev An experiment using Jev to generate landing pages in real time. Jev is great at picking between different options so I attempted to break down the langing page UI into a set of choices for the model to choose from. How this works:
1. Users type what kind of page they want and Jev decides what each line is (headline, fact, assertion etc.)
2. Jev makes choices for each layer of the page (opening, stage, light, depth etc.). Parameterizing design is hard (impossible?) but this took some tuning with Claude
3. A full page is constructed with a single call of about 90 questions. Of course using an LLM would be more prudent for step 1, but I wanted to check the limits of Jev's understanding (it works pretty well in my opinion!). https://ift.tt/za42MEB October 5, 2026 at 12:30AM
Show HN: I built a vacation optimizer and later found out tools like this exist https://ift.tt/KE7wye1
Show HN: I built a vacation optimizer and later found out tools like this exist https://ift.tt/nZQa7Us October 4, 2026 at 10:49PM
Show HN: French Revolutionary Calendar https://ift.tt/QXNxA6M
Show HN: French Revolutionary Calendar https://ift.tt/JHKc8v5 October 4, 2026 at 11:38PM
Saturday, October 3, 2026
Show HN: Revline – an AI web app that researches TikTok and Instagram for you https://ift.tt/lyn5bFG
Show HN: Revline – an AI web app that researches TikTok and Instagram for you I build revline because i was trying to figure out how viral content works across tiktok and instagram, since i am mainly developer, i was confused about how to create content that gets me users.
Actually going viral requires single founder ( me ) to upload 3-4 videos a day, brainstorming all the formats was not effective to do alone, so i build the agent that does all the scrolling for me. would appreciate feedback https://tryrevline.com/ October 4, 2026 at 01:20AM
Show HN: Automating my 35mm film scanning pipeline https://ift.tt/HvWE4lc
Show HN: Automating my 35mm film scanning pipeline https://ift.tt/cFlf4h9 October 4, 2026 at 12:35AM
Friday, October 2, 2026
Show HN: Our space game has a built-in RISC-V emulator that runs Linux https://ift.tt/dLXxV6f
Show HN: Our space game has a built-in RISC-V emulator that runs Linux We're a tiny indie studio, all with a background in the demoscene and for about 3 years now we're developing a space planet terraforming game called SEEDS - Echoes Beneath the Sands. Our protagonist Naxiah is stranded on a small, desolated planet. You're working for an intergalactic distributor of seeds and terraforming equipment, to kickstart new planets in far away galaxies. You must deliver seeds and utilities to an unchartered region, but on your way, you crash on a small planet. Luckily, you have some equipment with you in the ship, so you're able to spawn a base and survive. But for how long? And is the planet really deserted..? :) Well, of course not, there are aliens and other space creatures. And of course, there's ROBO RB-23, which accompanies your adventure. The game is written in Object Pascal and uses our own game engine PasVulkan, our own physics engine, as well as our own RISC-V 64-bit emulator PasRISCV. We love to build stuff, a lot of these things are FOSS. The emulator is quite complete (even with RVV, but that's way too slow emulated to be useful) and runs a stock kernel (6.18.3 as of today). We're using Alpine Linux for our base distribution. All our in-game programs to interact with the planet are native RISC-V Linux programs. We also have our own (free and open source) scripting language POCA (class and prototyped based, JS/Lua inspired, which makes it quite easy for us to script things like HUD, animal behaviour, our story flow graph, etc. We have most of the wiring ready for players to be able to program and reshape and populate the whole planet in automation, although it's not our primary concept for the game (building stuff by hand, making the planet beautiful is also fun). As it's a normal Linux system, you'll be able to use Ruby, Python, or even Rust, C, C++, and Go to change and automate your world. We're also thinking of making a bunch of small games which you can play inside the game, in the base on a computer, on an arcade machine, or in the spaceship. And of course, it also runs DOOM. We already prototyped a small 2D space shooter and a small arcade game, both playable within the game. We also feature a table with hologram games to be played, e.g. chess and a certain block game. We also plan to make this hologram game engine available from the computer, which means, you'll even be able to program own hologram style games. We generally want to make the in-game computer more hackable in future, so that people can create and share their own little games. Currently, it's only a sandbox game (with beginnings of automation), to build stuff with our pre-made items. You can also assemble these to other, ready-made and shareable bigger items. This is the foundation for us to further develop survival and complete our story mode. Our game is still under active development, and we finally got our Steam page ready. We have tons of other things planned, but right now, it's more about polishing and quality for the Early Access. You can watch some videos and screenshots on our website. And we'd love to read your thoughts about his. It has been quite a ride, so far.. https://againstallodds.games/ October 2, 2026 at 05:40PM
Show HN: Arda – Marketing agents to help brands win Google and AI Search https://ift.tt/pXUtArg
Show HN: Arda – Marketing agents to help brands win Google and AI Search https://www.ardainc.com October 3, 2026 at 12:48AM
Show HN: We built a real-time desktop companion for patient's telehealth visits https://ift.tt/BxKkoEu
Show HN: We built a real-time desktop companion for patient's telehealth visits Hey HN, Some friends and I built an app called Sidekick that joins your telehealth visits in real-time. It explains medical terms and suggests useful questions, including ones you may not have thought to ask. After the visit, it provides a summary including next steps, so you can focus on the conversation instead of taking notes. The idea of Sidekick came from our team's experiences with healthcare. Our founder, Rahul, had stage three cancer at 17. His mom had to manage everything around his care: taking notes through treatment, figuring out what questions to ask, etc. A decade later and he’s now in remission, but still sees a lot of specialists and deals with the same challenges. Another teammate became a caregiver after his dad had a severe stroke. He found himself taking notes, looking up medical terms, and trying to keep track of what was said during every doctor visit. Others have dealt with situations like this when navigating pancreatitis and other serious health issues. These experiences got us interested in how we might be able to reduce the burden on the patient or caregiver. A lot of products focus on simply recording or summarizing the visit. That's useful and a step in the right direction, but we kept coming back to the fact that a lot of the challenge originates while the conversation is still happening. If you can catch something in the moment, you can address it with your clinician immediately and potentially lead to better outcomes. If you realize you missed something after the visit, then you have to try to get another appointment (and in SF I often can't get another doctor visit till 2 months later). So we built Sidekick, to help patients and caregivers during a visit in real-time to translate medical jargon and suggest useful questions, ideally without becoming a distraction. If you want to try it, you can download it on our website. Sidekick is desktop-only for now, but we’re working on a mobile app for in-person visits. We would be especially interested to hear feedback on the live part of the experience. What are your thoughts on the balance between live help vs after-visit? Would something pre-visit be more helpful? https://openhand.health October 1, 2026 at 09:27PM
Thursday, October 1, 2026
Show HN: Open-source model routing for coding agents at Astra-level performance https://ift.tt/Nt8nf49
Show HN: Open-source model routing for coding agents at Astra-level performance A few months ago we started building a model router for coding agents because we thought we could outperform any single model with an ensemble approach. Recently we’ve achieved that milestone and I want to talk about how we did it. First of all, a quick explanation: the Weave Router ( https://ift.tt/1Bd0Cgu ) plugs into any coding agent (e.g. Claude Code or Codex) and intelligently switches between LLMs. So, for example, Astra handles tricky debugging or complex system design tasks, and Deepseek v4 Flash handles simple frontend updates. What we’re announcing today is our new routing model, which we’re calling Weave Router 2.0. We benchmarked 2.0 against GPT-6 Astra on Terminal Bench 4.0 and SWE Atlas. On both benchmarks, the router had equivalent pass rates. On Terminal Bench, the router hit 52% of Astra’s cost, and completed tasks 2.2x faster. On SWE Atlas, the router cost 54% as much as Astra and ran 2.5x faster. (Full results on our website at https://ift.tt/VugHikQ !) It turns out training a model to route effectively - taking into consideration model capabilities, costs, cache awareness, and more - is a really hard problem! I want to talk about three ways we were able to improve so much over the last few months: 1) a new architecture, 2) larger training data set size, and 3) smarter cache-eviction impact calculation. 1) a new architecture. Our initial approach used an RL model without many priors. While RL is still an important part of the story, the cost of fully exploring the space of routing decisions is very high, so we’ve taken some shortcuts that have significantly improved performance. Consider how large the search space for the routing problem is. Take a typical coding agent session, with ~100 agent turns (i.e. 100 LLM API calls). Technically there are 100 chances to select a model. If we assume a roster of ~10 models (of course there are lots more but we can remove any that are Pareto dominated), then there are 10^100 possible paths through that session. We simply cannot explore all of them! So that's why clever tricks to shrink this space are so important. In particular: we trained a hidden Markov model to trace the session state, then a classifier maps the session to one of a few buckets of similar models. Using the HMM allows us to evaluate not just where a session is currently, but how it got there . We've gotten significantly better performance on bucket selection by incorporating that information - we believe this is because two sessions that might look quite similar to a naive classifier are much better distinguished by this HMM approach. Using this HMM + classifier to select a bucket first significantly shrinks the space to explore, by throwing out most models that could not reasonably serve the given session. This rearchitecture was the single biggest performance unlock! 2) larger training data set size (much less technically interesting but still an important part of the story). By using frontier LLMs to help us label a larger and more diverse set of coding agent sessions, we were able to bootstrap the two models discussed in 1) to a better state, while also providing even richer reward signals for RL. 3) smarter cache-eviction impact calculation. One of the hardest parts of routing well (if you care about saving money) is using the model caches intelligently. We built a subsystem that can calculate the expected value of switching models (and thus paying a high one-time cost to fill up a different cache) much more accurately, helping us avoid costly and unnecessary switches in more cases, while still switching when the benefit outweighs the cost. This is where most of our improvement on cost has come from. We still have a lot of room to continue to improve (we won’t rest until we’re consistently beating Astra/Fable, not just tying!) but matching frontier model performance was a huge milestone for our routing model, and in my opinion validates our initial hypothesis that an ensemble of models can do better than any single model ever could. Our router is open source ( https://ift.tt/1Bd0Cgu ) so anyone can try it out. Or if you prefer you can use our hosted version ( https://ift.tt/VugHikQ ). September 30, 2026 at 11:58PM
Show HN: Gutsy, a 0.8B Jev-compatible decision model that runs on your CPU https://ift.tt/sAZMfLY
Show HN: Gutsy, a 0.8B Jev-compatible decision model that runs on your CPU https://ift.tt/xbHn8ep October 1, 2026 at 10:44PM
Subscribe to:
Posts (Atom)