news.ycombinator.com
Show HN: KRNL – a quieter browser for the essence of the webhttps://news.ycombinator.com/item?id=48869197Download, signup, pay? Never for this demo you can try now: ssh krnl.duetbrowser.com But what will you see? This: https://www.youtube.com/watch?v=C3lMIGjkwGMSelf-hosted available (+ needed BrowserBox): https://win9-5.com/krnlAdvantages:- set the typography you like (your terminal font, size, dimensions) and it just works - the web renders your way- light or dark mode just works- set colors you want with terminal color overrides- Supports: file uploads (disabled in the free demo for security), file downloads, JavaScript modals, HTTP Basic & Digest Auth, all modern JS, multiple tabs, "Open in new tab", back/forward/search from omnibox bar - all from terminal.If you self host - you also get all the benefits of BrowserBox, such as: GUI breakout for CloudFlare/CAPTCHA (just use your bbx login link, solve it, then back to term), WebAuthn (for now only if you're on macOS), remote DevTools (inspect any page), etc.Mostly tho this is not about that - it's about the text, the quiet, the world witJul 11, 2026 5:59 AM
github.com
Show HN: An MCP server that lets AI agents request disability insurance quoteshttps://github.com/seaworthy-io/seaworthy-mcpHi HN. I own an online insurance agency, a painfully old-school industry. I am not a developer. I vibe-coded it all in Claude Code, then reviewed and tested before anything shipped. Which is part of the story I guess, that a small firm with curious and reasonably capable people can now operate agent infrastructure that would have required a contractor two years ago.I'm trying this because a growing share of our prospects start by asking their AI of choice what coverage to buy and who to get it from. Assistants can read my site, so like everyone else I'm trying to get recommendations via content. But the next natural step is agents actually taking actions on sites rather than just reading them, so I created seven tools, six read-only (a verified-facts knowledge base, carrier comparison, specialty and education guides, benefit-cap math, rider definitions) and one action, "quote request," which writes a lead straight into our CRM through the same Salesforce web-to-lead path that our web fJul 10, 2026 9:56 PM
github.com
Show HN: 9lives – Self-healing test runner that refuses to mask real bugshttps://github.com/Quality-Max/9livesI built 9lives because coding agents kept breaking my Playwright tests in the dumbest way: rename a button, the test goes red, and the agent — or I — rewrites a perfectly good test.9l heal login.spec.ts runs the test, classifies the failure, and heals it in tiers:Tier 1 is offline and deterministic. When a selector stops matching, 9lives re-finds the element in the page snapshot Playwright captures at failure (data-testid > id > aria-label > text > class — the most stable surviving anchor wins) and rewrites the locator. No LLM, no network, no account. Most selector drift heals in seconds, for free. Unlike Healenium-style tools, it needs no baseline DOM from previous green runs — it works from the failure itself.Tier 2 uses the subscription you already pay for. For structural changes it shells out to your installed claude / codex / opencode CLI in headless mode, so the fix rides your existing coding-agent plan. No new API key to mint. (Raw ANTHROPIC_API_KEY / OPENAI_API_KEY work too.)EvJul 10, 2026 6:01 PM
github.com
FreedomLang – native x86-64 HTTP with no libc / no CRThttps://github.com/DO-SAY-GO/freelang/blob/main/examples/HTTPD.mdI'm sharing a small milestone today - a tiny HTTP server written in FreedomLang (freelang), which is a new, experimental, AOT-compiled language I've been working on.The demo (httpd.flx) compiles down to native x86-64, and the resulting server runs with no C runtime/libc dependency (it uses direct syscalls on macOS and Linux, and Winsock plus Kernel32/WS2_32 on Windows).If you want to poke around you can clone the repo, build it (the compiler is just a Node.js script, very inspectable), and test it with a simple curl localhost/health or curl localhost/ . You can also inspect the native machine code with tools like otool -L on macOS to verify the lack of Mach-O dynamic-library dependencies.The reason I chose a tiny HTTP server was it shows an actual socket lifecycle (accept/read/respond/close), but it doesn't pretend to be production-grade.If you look at examples/httpd.flx[0] you can ysee how eveyrthing works - and how the language models the failure states. Why build a new language? TheJul 10, 2026 4:25 PM
sxp.studio
Show HN: SubjectiveZero, an open-source agentic node editor for creative codinghttps://sxp.studio/apps/subzHey there,My name is Clem, I've been a solo indie dev for a couple years now, exploring frontier tech like XR and agentic workflows in the context of creative / interactive work.I've been building creation tools for a while and some common design challenge is to figure out the right level of abstraction for your tool. You can always make it super advanced and complex with low level concepts (shader composition, actual code etc.) but then you get something with a high complexity / learning curve. On the other hand, if you make your tool too high level, it might be easier to use at first, but people will most likely hit a wall eventually and start fighting with your tool to get their edge case done (you see that on mobile a lot actually).With this prototype (called SubjectiveZero), I'd like to imagine that we can kind of move the "slider" on the abstraction layer, meaning that you can actually start with prompts that describe the goal, and you can go as high level (stay with abstract proJul 10, 2026 3:23 PM
mojavepaint.app
Show HN: Mojave Paint for macOS, edit images like it's 1999https://mojavepaint.app/We all need a swiss army knife for images. A tool for really getting in there and dealing with the pixels precisely. You might be combining images into a sprite sheet, you might be matting images to squares, cutting out solid backgrounds, re-tinting graphical elements. For a website, a video game, for an app.There's a million photo editors, there's a million design tools, and Mojave Paint can do those things in varying degrees but "graphical asset production" is its first and primary mission.It's cheap! $9.99 (one-time, sad I even have to say that) for the "Pro" version. It's Mac-only, and Apple Silicon Mac at that. Does it look a little like Photoshop 5.5 from the year 1999? Maybe a little but also it's got plenty of its own original ideas. (Blog post on that is forthcoming…)I've always been passionate about image editing tools for the technically-minded so I'm throwing my hat in the ring in that space. And sure GIMP is for the technically-minded but boy is it ugly on the Mac!Jul 10, 2026 12:29 PM
bing.com
Newsletters Are All the Rage. As an Author, Here’s What to Use Based on Your Needshttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8ac5add0154303a78b4726951ac5ad&url=https%3A%2F%2Fwww.cnet.com%2Ftech%2Fservices-and-software%2Fbest-newsletter-platforms%2F&c=5184381445036580967&mkt=en-usNewsletters are some of the most powerful marketing tools around. While a social media post might reach a handful of your followers, newsletters are sent to every subscriber. This makes a newsletter ...Jul 10, 2026 10:00 AM
sbnyc.app
Show HN: I built a free app for New Yorkers to save money on grocerieshttps://sbnyc.app/I built this because I see that grocery savings are achievable in NYC. People usually just go to the store they're used to going to, and it's rarely worth the effort of combing through card cashback, weekly coupons, CPG rebates.Most people leave real money on the table by not stacking them, and even more don't even know that these deals are out there.... so I built a way to automate it.You can use it for free, no login, currently NYC-only with ~690 stores.I built it so that you just search whatever you want (use commas if you want to search multiple items). Or - use the AI tool to help shop for you. If you're curious, it's powered by a trained LLama model.Honest limitations are coverage and freshness. Id love some feedback on where the data looks wrong or is stale.Question for the room - what to prioritize if you're working with messy, multi-source retail/pricing data? Is freshness or coverage the top priority if you cant get a uniform response from every source? curious on what to priJul 10, 2026 12:17 AM
ello.com
Building a real-time AI tutor for 5-year-oldshttps://www.ello.com/blog/teaching-a-child-in-1000-msHey HN! We've spent the good part of this past year building an AI tutor that teaches kids ages 4-9 reading, math, ESL and more. Getting an AI tutor to effectively teach a child turns out to be a really hard technical challenge, this took getting the underlying architecture right.Our tutor steers the UX in real-time and makes complex decisions on the fly. Doing both at conversation speed required us to replace the standard tool-use loop. We built our own tutor harness that utilizes a streaming interpreter that executes actions, while an asynchronous planner model reasons ahead of the conversation and makes calls that drive the child's learning. On top of it all, we developed a safety system that checks every turn without it causing an interruption to the activity and conversation flow.Effective teaching isn't just about answering a child's question quickly, rather making the right move at the right moment. AI is also going to be an integral part shaping how this generation of kids learJul 9, 2026 8:51 PM
kastra.ai
Show HN: Policy enforcement for Claude Code, Cursor, and Codexhttps://kastra.ai/Show HN: Runtime authorization for Claude Code, Cursor, and CodexHi HN, Fernando and I built Kastra. Kastra intercepts AI agent tool calls and evaluates them against deterministic policies before they execute. This is aimed at developers using coding agents like Claude Code, Codex, Cursor, and OpenClaw.We built Kastra after one of our Cursor agents almost executed DELETE FROM customers WHERE status='test' against a production database. We caught it before it ran, but it made us realize that nothing in our stack actually decided what the agent was allowed to do. What mattered for us wasn't the mistake; it was realizing nothing in our setup would have stopped it if we weren't actively on top of it. LLMs are probabilistic, and prompts influence behavior, but they don't deterministically decide what an agent is allowed to do. Without a deterministic policy system, nothing could have decided what it was allowed to do.Kastra pushes an allow, hold, and deny decision before the action runs. YoJul 9, 2026 3:26 PM
news.ycombinator.com
How to feel safe delegating to AI agents?https://news.ycombinator.com/item?id=48846739I feel more and more that I am loosing control of what's happening under the hood. Agents move fast and chaotically. Are there frameworks, tools or governance solutions to give us back more control?Jul 9, 2026 2:43 PM
github.com
Show HN: Ffvii Realtime – speed up FFVII Tactical Mode slow-mo to realtimehttps://github.com/robesris/ffvii-realtimeI made a tool that detects and speeds up Tactical Mode slow-motion in Final Fantasy VII Remake/Rebirth/Revelation so that it plays smoothly in real time.Motivation: The Final Fantasy VII Remake and Rebirth games are beautiful and have a unique and interesting combat system. However, what's fun to play is not necessarily fun to watch, and FFVII combat footage is filled with frustrating spans of slow motion while the player is choosing a command in Tactical Mode.This tool detects Tactical Mode segments and speeds up those segments 100x, which, at default settings, equates to real time.What about the audio? In the games, the audio continues to play at normal speed, so speeding it up by 100x just creates a jarring "gap" in most cases. By default, ffvii-realtime smooths over the audio so well that in most cases the artifice is not even noticeable. This is probably the most interesting and non-intuitive part of the project, and it's covered in great detail in the README: https://github.com/rJul 9, 2026 12:24 PM
news.ycombinator.com
I am tired because of today's condition of project markethttps://news.ycombinator.com/item?id=48844537I have been posting everywhere Hacker news, reddit , devto every possible platform. I am Raj shekhar, a fourth year college student , I have been working on a project for the last 7 months, like i had a huge problem that i tried solving and while solving the problem i realized that this can be turn into a huge Dev tool that everyone can use , so i made it work like everyday almost while having the semester exams in between (for context i am a electrical engineering student and I dont even like electrical, I am a tech lover) so I have to focus on the exams also, while i was building it, I never though it will be become this complex and its pretty helpful also it become a huge thing.Now when it come to post the project on the internet every kid in his home making some cheap vibe coded app/website, the market is filled with garbage projects written by ai, I am not saying that i havent used any AI but I have used it to debug the code and cleaning the repo and for readme. But my point is thJul 9, 2026 12:03 PM
linear.gratis
Show HN: Free and OSS way to extend client-facing functionality of Linearhttps://linear.gratis/Hi HN, I've been working on building this tool for myself and my software dev shop so I can easily share project specific + interactive views with my clients.Have been slowly adding/improving the platform over the last 9 months and while I'm sure there are still some gaps, it's at a point where it would be nice to share with the community.As per the domain, the intention for this is to always be 100% free with MIT license.Current plan is to try and polish the UI to the level of Linear + try and replicate the sync engine experience (as close as possible noting that Linear API is the main bottleneck).Any feedback is welcome either on here or as an issue :)Jul 9, 2026 3:20 AM
github.com
Show HN: Moo, Git versions code, moo versions machineshttps://github.com/heyito/mooI'm Evan, CTO of Ito (ito.ai). We do code review that actually runs your code. When we're developing locally, we run lots of coding agents in parallel against the same repo, and we kept hitting the same wall: git worktrees isolate files, but the database, ports, and services all collide.Moo is a brand new internal tool we use to moove faster and thought we'd share it.Every branch or agent attempt gets its own isolated Linux machine, and the machine's state is versioned with your repo. Moo save snapshots the runtime tagged with the current commit. Check out an old SHA, run moo new, and you boot the exact runtime that existed there.It's microVMs with copy-on-write (CoW) disks under the hood. Forking a 20 GB environment isJul 8, 2026 6:22 PM
tmahmood.github.io
Show HN: Fluffy Sparrow – Frictionless, Fast, Note Taking Apphttps://tmahmood.github.io/fluffy_sparrow/Hello HN,Taking notes should be frictionless, but for me with most of the note apps, there are too much frictions to quickly put down what I have in my mind, before long forgetting what I was intending to writeSo, for that I built Fluffy Sparrow, a frictionless, fast, simple note taking appI have been working on it for last two years, on and off. I abandoned first version because it strayed away from the concept that I had in mind.The primary target of this app was to be frictionless, and fast, so that I can just fire it up, and put down what's in my mind without thinking, and store all data in simple text format, which can also be filtered using *nix tools.And, happy to say this attempt clicked.It supports, besides taking quick one liner notes; to-do, project, tags, inline calculations using fend[1], filtering, search, contextsAnd more features, like time tracking, reminders, KanBan, saving links, scripting are also in the pipeline. BUT you need not to use any of these features, if yoJul 8, 2026 5:05 PM
ghoster.app
Show HN: Ghoster – Crowdsourced database of which copanies ghost job applicantshttps://ghoster.appI started building ghoster in 2021 after trying to switch a career and noticed the issues with ghosting in the HR world were extreme (and still are).In the past 8 weeks I applied to 359 jobs of which 175 ghosted me!Ghoster connects to Gmail (Read-only), scans only emails from emails you applied to and automatically updates your application statuses to ghosted, rejected, interview scheduled, offer. Nothing is stored except the detected status.The tracker is always free and unlimited. The Pro tools ($14.99)/mo) are where it gets more interesting.Resume tailoring - paste a job description and get your resume rewritten for that specific role . It shows the ATS keyword score and job fit score before and after processing so you can see what changed and why.Cover letter generator - produces a tailored cover letter with all your info, no filler phrases, no randomness and matches the resume.Interview prep - this is my and users favorite. Generates a full prep for a specific role strenghts and gJul 8, 2026 2:09 PM
news.ycombinator.com
Ask HN: Why aren't we collaborating on the prompts we give to our AI agents?https://news.ycombinator.com/item?id=48831425Now that code is so cheap, I feel like the real work is the spec, the context, the plan (aka the prompt) you hand over to the AI agent. That's where you decide what's actually getting built and how.That step still seems to be completely individual. You write it alone, the plan you formulate is local and private, and your coworkers only see the result when the PR is opened. On a team, that's the exact moment alignment should happen, and it's the one place we don't work together nowadays.GitHub Next recently showed a prototype that gets close to this collaborative experience: https://maggieappleton.com/zero-alignmentI also wonder how other knowledge workers work and how they review each others' work. I believe they would also benefit from a "collaborative prompting" environment, where everyone sees what the others are working on and can step in at any point.Why isn't this normal yet? Missing tooling, cultural habit, or does collaborating on a prompt not help at all?Jul 8, 2026 1:04 PM
github.com
Show HN: CLRK, an open-source agent runtime with gVisor and MitM guardrailshttps://github.com/apoxy-dev/clrkTL;DR: we built a framework-agnostic agent runtime that uses gVisor for isolation and runs on k8s. It’s open-source under AGPLv3Recently we’ve been working on a customer support “AI assistant” - essentially an interactive knowledge base/L1 support but with an option to touch resources that belong to a customer it’s talking to. We found existing tools to be lacking in these aspects:1. Fully intercepted i/o. We wanted to trace out LLM calls as well as any other networking calls attempted by the harness so that guardrails and audit trails apply to all current and future systems uniformly. Nobody’s agent can accidentally make raw database calls or send PII data to an overseas LLM provider.2. Coherent API and framework agnostic. There’re a lot of frameworks out there that do similar things in slightly different way and most we found had telemetry, guardrails and other tooling tightly bound into the framework. We wanted something with an infrastructure-first approach because we think it’s a Jul 7, 2026 7:45 PM
github.com
Show HN: Bike4Mind – open-core AI workbench; any model, agents, RAG, self-hosthttps://github.com/Bike4Mind/bike4mindHi HN, I am Erik Bethke, game dev that has been making online games for 30+ years (Starfleet Command, GoPets, FarmVille, Mafia Wars).For the last 3 years I have been building Bike4Mind and now I have to put my big boy pants on and 'go to market'. Ugh, it would be so much more cozy to fire up a few more agents and fix a few more bugs, polish some more UX - or hey how about a new feature that does Q, R or S?What is Bike4Mind? I started Bike4Mind to be the home of a bunch of AI augmented game development tools, and over time have just kept adding features. It is a full cognitive workbench we have all of the features - notebooks, files, data lakes & RAG, agents (different forms), generate share-able artifacts in react or python. We have secure websockets to connect to our CLI - B4M CLI, so it is truly batteries included - and runs any model - OpenAI, Anthropic, Gemini, xAI, Bedrock. Switch models during a conversation, fork a notebook, run 4 different models on the same prompt, loads of imJul 7, 2026 7:17 PM
