github.com
Show HN: Rudoc – a 4.5MB Rust document converterhttps://github.com/asong56/rudocI built Rudoc because I wanted a smaller tool for simple file convertion workflow. (Pandoc was so heavy for me) I once wanted to try Pandoc, but after I downloaded it from GitHub, I found that it is around 70 MB. It also needed a lot of dependencies.Rudoc is not a substitue of Pandoc, it supports limited set of formats (only txt, md, html, typ, docx, pdf, pptx, xml, json) and doesn't support advanced formatting. The most common workflow is just turn a markdown file to a html. PDF generation can work without Typst, but PDF input conversion requires Typst as a backend. It has no other runtime dependencies and can run as a single binary, by the way.It is pretty easy to use, just use the format of rudoc input.file output.file One sample usage is like this: ```powershell rudoc OhPDF_ROADMAP.md.txt r.md [rudoc] OhPDF_ROADMAP.md.txt → r.md (txt→md) [rudoc] done in 52ms (6286 bytes) rudoc r.md r.html [rudoc] r.md → r.html (md→html) [rudoc] done in 4ms (7363 bytes) ```Some complex formats (suchJul 25, 2026 10:11 PM
news.ycombinator.com
The Cost of Seamlessnesshttps://news.ycombinator.com/item?id=49048352The Hidden Cost of Frictionless Technology Taotuner — 2026I recently asked an AI assistant for restaurant recommendations. It listed five places, complete with ratings, distance, and popular dishes. I picked one, went, had a fine meal. Later, when a friend asked what I'd eaten, I couldn't remember the name of the restaurant.The experience wasn't bad. It was frictionless. And it left almost no trace.We've built tools that eliminate the gap between want and satisfaction, question and answer, curiosity and resolution. We call it seamlessness. We've assumed it's progress. But the gap wasn't inefficiency — it was where memory, judgment, and understanding used to form.What if removing it is making us less capable?The Pattern This isn't just about restaurant recommendations. The same dynamic shows up across AI, therapy, cities, and education.Take AI assistants. The industry evaluates them on speed, accuracy, and satisfaction. Those metrics capture whether the tool works. They don't capture whJul 25, 2026 3:24 PM
news.ycombinator.com
Why My Open-Source Project Hasn't Done Betterhttps://news.ycombinator.com/item?id=49046999Since the beginning of 2026, many projects such as RTK, Caveman, and Ponytail have claimed that they can reduce token usage by 80–90%. Some of them gained tens of thousands of GitHub stars in a very short time.My previous startup was building an e-commerce chatbot. In fact, the product was almost identical to the Universal Commerce Protocol (UCP). After UCP was released, I decided not to continue pursuing that direction.Earlier this year, I started turning the architecture we had built for e-commerce chatbot agents into an SDK. Put simply, its core idea is to use state machines and deterministic execution sequences to reduce the number of round trips to LLM providers by around 80% for the same task, thereby significantly reducing token consumption.Here is the problem: I understand very clearly why tools such as RTK are ineffective in genuinely long-running tasks.I published a report on July 18, and JetBrains published a report on July 20 that reached essentially the same conclusion: toJul 25, 2026 12:23 PM
news.ycombinator.com
Ask HN: How would you harden AI changes to a 1M-line legacy SaaS before review?https://news.ycombinator.com/item?id=49045271I’m not a software engineer, but I’ve been running an experiment to see whether agentic development could produce a useful prototype on top of an existing SaaS codebase.The codebase is 1M+ lines, 15 years old, hosted on Azure, and primarily written in C# and React.The prototype needs to be available for customer testing in September. No developers were available to work on it full-time, although I could occasionally get help with specific technical issues. An engineer will evaluate the implementation in August and decide how much confidence we can have in the AI-generated code so that we can decide how to “convert” the code to production-grade.My question is what I can do before August, primarily using AI tools and manual testing, to make the code as robust and reviewable as possible. I want to increase the likelihood that the AI-generated code would be so good that the path to production-grade would be closer to “copy-paste” than building everything again from scratch.Environment The Jul 25, 2026 7:19 AM
github.com
Show HN: WhipDesk – Control your full dev machine from your phonehttps://github.com/BinaryBananaLLC/WhipDeskHi HN, I built WhipDesk, an open-source, mobile-friendly remote desktop tool for controlling your full development machine and managing AI coding agents from your phone’s browser.It gives you access to the entire desktop while also adding workflows for monitoring and managing tools such as Claude Code, Codex, GitHub Copilot, and othersThe two questions I hear most are: How is it secured, and how is it different from existing tools?Security Following the modern approach taken by open-source remote access tools such as RustDesk, WhipDesk is fully open source, so its client, agent, protocol, and security model can be inspected rather than trusted as a black box.Local connections can run entirely within your network without an account. Remote connections use end-to-end encrypted WebRTC. Most connections are peer-to-peer; when a TURN relay is required, it only forwards encrypted traffic. Each connection also requires the machine’s access PIN as an additional security layer. More details areJul 25, 2026 2:49 AM
news.ycombinator.com
Show HN: How well do you use Claude Code?https://news.ycombinator.com/item?id=49042653A bunch of companies that I spoke to had their own claude & codex OTel dashboards that showed spend + seats per month.However, none of the dashboards actually analyzed how the engineers worked with the tools and if there were any areas for improvement! That's why I created https://www.promptster.ai.Managers get aggregate level view of code quality and how that ties with team workflows (nothing on a per-engineer level). While engineers get personalized coaching on how they can save tokens while keeping output high.We also have a tool built for individuals to test their local claude code setups. It's open-source and completely local. You can check that out here: https://github.com/pa-arth/cc-audit.Would love feedback on either and reach out if you're interested in purchasing the teams product!Jul 24, 2026 11:00 PM
tripverdict.ca
Show HN: TripVerdict, in 90 SEC know whether your next trip fits the budgethttps://www.tripverdict.caI was planning a trip to St Johns NewFoundland with my family, while researching on multiple websites such as skyscanner.ca, booking.com, expedia, airbnb, etc for 3 hours, we realized the total trip cost is over $2000.This triggered TripVerdict, as I work in TravelTech and knows most of backend of the system, I realized I can build pre-research gut check tool which help people know what fits vs what not to save their precious hours.This product solve 3 pain points: 1. I know my budget and destination, does it fit? 2. I know my budget and confused between 2-5 destinations, how do they compare? 3. I know my budget and want to which destination I could makeUses real time flight, hotel and rentals prices to provide cheap vs comfortable options. Once you are satisfied with the verdict, then booking links are shown to proceed.There is no tool out there which provide traveller with this opportunity, every OTA is trying to sell. We are here to protect your budget.Tech Stack : Typescript, Node.Jul 24, 2026 7:37 PM
github.com
Show HN: Poddie – A local-first, text-based video and podcast editorhttps://github.com/SinanTang/poddieHi HN, I've been itching to start a podcast this year and needed a way to edit the footage quickly. I'm too cheap to pay for Descript, (and as a dev with tendency to build tools instead of working on the real task) so I just built my own local version of Descript which I call Poddie.Poddie does it e2e for free: it generates transcript using local Whisper model, lets you edit by reading through the transcript and deleting words, auto-trims silence, offers video/audio/captions export with optional caption burn-in. You can also do waveform editing for more precise cuts if desired.I'm open sourcing it so other creators could try it out and use it for their own workflows.Happy to hear any feedback!Jul 24, 2026 4:28 PM
news.ycombinator.com
Ask HN: How much software do you build for yourself these days?https://news.ycombinator.com/item?id=49035758With the cost of building software continuing to drop, what have you found yourself doing more: using other people's software, or building it yourself?I was thinking about this... I've built our family's financial system (it started as a simple budgeting tool, but now it tracks everything), a Mac app to create and edit videos, quite a few developer tools for personal use, a bunch of apps for my kids, some games, and so forth.All of these already have plenty of commercial offerings in the market, and some have open-source alternatives, but the level of personalization I get by building them myself beats any argument (IMHO)!Interested to see what others are doing. :)PS: I've been writing software for 8 years (5 of them professionally).Jul 24, 2026 1:58 PM
news.ycombinator.com
Anyone else like llms.txt for architecture docs?https://news.ycombinator.com/item?id=49035043I'm a fan of using the llms.txt pattern of progressive disclosure for internal app documentation. Agents just produce more consistent code with these.Two things I regularly use it for are documenting decision records, under app/docs/adr, and to manage AWS learnings (https://github.com/jbdamask/aws-learnings-library).For example, in the case of ADRs, as the codebase grows bash tools aren't enough for the agent to get the big picture. If a new feature requires async processing, the agent can orient itself by checking the adr llms.txt, first, then designing according to project standards, e.g.llms.txt - [0002 Distributed SQS+Lambda Worker Pipeline Over Single-Container Agent]: [Accepted] Rejected porting the local single-session agent pattern to AWS; Lambda time limits, the propose-review gate, and idle-cost requirements each independently require the distributed worker model.0002-distributed-worker-pipeline-over-single-agent-lambda.md "Each ingest job is a message on an SQS queue. A LambJul 24, 2026 1:11 PM
bing.com
Will Medicare Pay for a Lift Chair? 2026 Coverage Ruleshttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8d20e271684530bf62fd649596d4b8&url=https%3A%2F%2Fhealth.usnews.com%2Fmedicare%2Farticles%2Fdoes-medicare-cover-lift-chairs&c=10747220040975395172&mkt=en-usLift chairs can be a safe solution for people who struggle to stand up or sit down with ease. They can provide a helpful tool to help individuals aging in place maintain independence. Here's how to ...Jul 24, 2026 8:39 AM
glossarypage.com
Show HN: GlossaryPage – Turn customer questions into knowledge pageshttps://glossarypage.com/Hey HN, I built GlossaryPage after noticing companies have long-tail content opportunities that never get created because they fall between blog posts and documentation.The primary use case is marketing teams building glossaries and learning hubs around these topics. The same workflow also works for support teams creating help centers.GlossaryPage helps teams identify unanswered topics, generate drafts with AI, review content, and publish structured knowledge pages.Here is a demo site: https://glossary.glossarypage.comSome things I have been thinking about while building this:1. How to avoid creating more low-quality content?This is a challenge for anyone building AI content tools, and is something I'm working on. The product is focused on generating drafts, but users are responsible for reviewing and approving what gets published. My guess is that as AI systems improve at evaluating quality, there will be less value in publishing low-quality content.2. Why not just use an existing CMSJul 24, 2026 3:44 AM
github.com
Show HN: Remortgaging with Claude, so I built a MCP for Bank of Englandhttps://github.com/moureauf/boe-mcpThe time had come to renew my mortgage, while exploring different scenarios I found myself keeping asking Claude Code the same questions. "What's the base rate", "when's the next MPC meeting", "how has it moved over the last year" etc. Doing that by hand a few times made me wonder if an MCP server wouldn't be a good fit for this, so I built one. Partly to solve the actual problem and partly because I hadn't built an MCP server before and wanted to try it out.It wraps the Bank of England's public (no auth or API key)IADB statistical database.Tools for the current base rate, rate history/stats over any range, the next MPC meeting date, and generic access to any IADB series by code (SONIA, exchange rates, quoted mortgage rates).Any feedback welcome!Jul 23, 2026 3:49 PM
github.com
Show HN: Palmier Pro – Open-source macOS video editor built for AIhttps://github.com/palmier-io/palmier-proHi HN, we are Marcos and Harrison, cofounders of Palmier (https://palmier.io). We are building Palmier Pro, an open source macOS video editor, with built-in AI generation and a local MCP server that connects to your agent. Here are a few demos:- Making some AI transitions: https://www.youtube.com/watch?v=hbM_-eR1GX4- Multicam editing with Codex: https://www.youtube.com/watch?v=SjS2q2LT1q8- Cutting long form clips into shorts: https://www.youtube.com/watch?v=PR66eN2ouuQWe built Palmier Pro as an internal tool when we were making AI launch videos for other startups. The main problem it solved in the beginning was the back-and-forth between AI generation platform and video editor. The iteration loop was awkward: AI videos → download → import to editor → edit → realize we need to change the AI video → repeat. So we built a minimal video editor where we could let Claude generate AI videos inside the editor.As we gave more and more tools to the agent, we wanted to push to see what else agentJul 23, 2026 3:11 PM
bing.com
Mavericks Have Not Received Lakers Trade Interest For Two Ex-Luka Doncic Teammateshttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8d63b934e34f3db1e9ecd7dcd36210&url=https%3A%2F%2Fwww.si.com%2Fnba%2Fmavericks%2Fonsi%2Fdallas-mavericks-have-not-received-los-angeles-lakers-trade-interest-two-ex-luka-doncic-teammates-&c=11654075784810737849&mkt=en-usAs the Dallas Mavericks look to retool the team around Rookie of the Year Cooper Flagg, they're on a mission to find the best-fitting talent possible. Because Flagg isn't the greatest shooter yet, ...Jul 23, 2026 9:18 AM
news.google.com
Worx WX533L 40V Circular Saw: Serious Performance, Unbeatable Valuehttps://news.google.com/rss/articles/CBMiekFVX3lxTFBlVVI4OXd0ekVwdVhBUVBCZ3Rqdk5ldnE1TllzcU15VEx0a3J0Qy1ublltVHlIOFFYSXFUN1JPODMxN3NhQjVNT2xjaS1uSENqMm9RLU8zdFdLQkt0NkRIX0h5YkhjcFJ6UDJzQmRXUXk1ejR0bDcxeDhR?oc=5Worx WX533L 40V Circular Saw: Serious Performance, Unbeatable Value Pro Tool ReviewsJul 23, 2026 7:00 AM
github.com
Show HN: Hanesu – An experimental workflow layer for AI coding agentshttps://github.com/jezmn/hanesuA few months ago, I became interested in Harness Engineering and started researching it. I realized that, in my experience, as tasks start to grow, agents end up losing context, repeating steps, or trying to solve the same problem multiple times. I wanted to build this workflow layer to provide a clearer structure for working on features, bug fixes, and patches, while persisting task state. Of course, Hanesu can make mistakes, but having a record of what was done helps a great deal in making it more likely that a second iteration will achieve the expected result.So why not just use CLAUDE.md and put all the context there? Because CLAUDE.md describes how to work, but it doesn't save the execution state of a task. Context is limited, and as a conversation grows, it becomes more difficult to maintain the entire state of a task solely within the chat.In addition, by using multiple agents, each one focuses on a specific stage of the workflow. Hanesu leverages existing tools (Codex, Claude CJul 23, 2026 6:58 AM
oriflow.in
Show HN: Oriflow - A Vibe Coding Platform for Building Personal Appshttps://oriflow.in/Oriflow is a vibe coding platform for building and customizing personal apps. It addresses three points that a traditional app or an app store inherently do not solve.- Personal: Users can create and customise precisely the apps that they need, with natural language instructions (vibe coding)- In Sync: The apps apapt to user's changing routines, priorities and circumstances.- Persistent Context: Apps in the platform talk to each other, so users dont have to botehr about context switch between fragmented apps.Oriflow is inspired by experiences shared by couple of our friends, who are neurodivergent, especially in terms of their challenges with finding apps that they can stick to. Two such sample app developed using Oriflow can be found inBrain Dump tool: https://oriflow.in/app/brain-dumpTool inspired by Wall of Awful by Brendan Mahan: https://oriflow.in/app/wall-starterPlease share your feedback and suggestions. We just started and we use a schema driven app development approach. Our scJul 23, 2026 3:21 AM
github.com
Show HN: Stream Android Emulators to a Browser Windowhttps://github.com/hsandhu/serve-avdThe npx serve of Android Emulators.Host your emulator for use with Agent tools like Codex, Cursor, or Claude Desktop — locally, over your LAN, or host on a remote machine and tunnel anywhere.Jul 22, 2026 9:30 PM
github.com
Show HN: Millwright – Rust-based, self-hosted LLM routerhttps://github.com/Northwood-Systems/millwrightHey HN,With the news of OpenRouter possibly being acquired and proliferation of hosted LLM routers (i.e. Ramp Router, Vercel’s AI Gateway), I saw the need for a self hosted solution focused on cost savings, transparency, and performance. So, I built an open sourced router with a simple CLI interface that can easily sit between coding agents and GenAI workloads.For the curious and lazy, at the moment, Millwright has the tools for,- Providers: OpenAI-compatible APIs, Anthropic, Amazon Bedrock- Routing: policy-controlled model roles (cheap, mid, frontier), cheapest healthy route selection- Protocols: OpenAI Chat Completions, Anthropic Messages, text and tool translation- Cache Affinity: role-scoped session lanes without serializing concurrent agent traffic- Spend Tracking: per-team costs, cache usage, model/provider mix, request traces- Cost Analysis: measured usage and modeled candidate economics (HTML, Markdown, JSON)- Reliability: bounded failover, circuit breakers, timeouts, concurrenJul 22, 2026 7:03 PM
