Directory

Companies tracked by Rails Radar

Hiring signals with observed stack evidence. Showing 1057–1104 of 1479 · page 23 of 31.

• RelateAnything 0 open listings

  discussion=https://news.ycombinator.com/item?id=49713313, title=Show HN: RelateAnything – Real-time open-vocabulary relation prediction

• Relay 0 open listings

  discussion=https://news.ycombinator.com/item?id=49642712, title=Show HN: Relay – a voice layer for dense dashboards (MIT)

• Remarc 0 open listings

  discussion=https://news.ycombinator.com/item?id=49615660, title=Show HN: Remarc – better contextual feedback for AI agents, Like any side project, Remarc started with my own problem: there was no good way to give AI coding agents contextual feedback on their own output.

• Remind Cards 0 open listings

  discussion=https://news.ycombinator.com/item?id=49725821, title=Show HN: Remind Cards – an offline-first flashcard app with AI generation, Hi HN, I’m a software engineer, and for the past several months I’ve been building Remind Cards, a flashcard app I originally made because I wanted a better way to study Estonian vocabulary. https://remindcardsai.com/app What started as a small prototype gradually became a much bigger side project. The app now supports: * AI-generated flashcards from any topic * Uploading photos of handwritten notes and turning them into flashcards * Automatically generating example sentences for your cards * Pronunciation and text-to-speech, so you can hear the words and example sentences * Custom decks and sub-decks * A 7-box Leitner spaced-repetition system * Multiple study modes: multiple choice, typing, matching, writing, and listening * Images for flashcards * CSV import/export * Offline studying * XP, streaks, achievements, and leaderboards Most of the learning functionality works locally. I mainly use the backend for authentication, synchronization, subscriptions, and shared/community features. One thing I deliberately avoided was turning the app into just an “AI wrapper.” AI helps with getting learning material into the app — for example, you can photograph your handwritten notes and generate flashcards from them, or have it create example sentences — but the actual learning experience is based around active recall and spaced repetition. I’ve built the whole thing as a side project alongside my full-time job. The first prototype took only a couple of days, but I’ve now spent hundreds of hours polishing all the small things around it. I’d especially appreciate feedback on: 1. The study/review workflow 2. Whether the app feels too feature-heavy 3. The spaced-repetition approach 4. The handwritten-notes → flashcards work

• Remoc 0 open listings

  discussion=https://news.ycombinator.com/item?id=49713330, title=Show HN: Remoc – Rust RPC where channels are first-class values

• Remoot, run Cursor Agent and Copilot CLI from your phone 0 open listings

  discussion=https://news.ycombinator.com/item?id=49624468, title=Show HN: Remoot, run Cursor Agent and Copilot CLI from your phone

• Rendly 0 open listings

  discussion=https://news.ycombinator.com/item?id=49683520, title=Show HN: Rendly – Build demo videos from code in seconds

• Rent vs. buy calculator that shows two different break-even years 0 open listings

  discussion=https://news.ycombinator.com/item?id=49701971, title=Show HN: Rent vs. buy calculator that shows two different break-even years

• Replaces the X-axis and timeline visualization in videos 0 open listings

  discussion=https://news.ycombinator.com/item?id=49648182, title=Show HN: Replaces the X-axis and timeline visualization in videos

• Replay 0 open listings

  discussion=https://news.ycombinator.com/item?id=49691257, title=Show HN: Replay – Audit silent prompt cache misses in AI agent transcripts, I built Replay because AI agent bills spike without a single warning or error in your logs when a prompt cache breaks. Whether your team runs Claude Code, Cursor, or Aider, the provider just silently re-bills your entire conversation history at cold-write prices instead of cached read prices. Replay is a zero-telemetry, zero-dependency local Go CLI that forensically reads your on-disk log transcripts, replays them against reverse-engineered provider caching state-machines, and names the exact turn the cache broke, the cause, and the exact token leakage. Repo: https://github.com/RedRobotKK/Replay Core Subcommands (What it actually does): - `replay` : Computes total avoidable re-billed cost across your local transcripts. - `replay diff ` : Pinpoints the exact turn the cache sharded and why. - `replay advise ` : Outputs structural prompt optimization adjustments based on your history. - `replay context ` : Ranks and isolates exactly what content is bloat-filling your context window. - `replay route --to ` : Simulates what switching models would cost, factoring cold-start prefixes. Why Go / Design Constraints: 1. Zero Dependency: Avoids messing with Python virtual environments or Docker runtimes for local diagnostics. 2. 100% Sandboxed: Absolutely no remote tracking, cloud accounts, or telemetry. Your prompt text never leaves your machine. 3. Mutant Testing Isolation: Keeps 76 real past defects frozen as re-runnable mutants to enforce code stability. It's completely free and source-available (BUSL 1.1). To check your local developer logs or audit the setup script before running, use: curl -fsSL https://replay.doctor | less

• Replay the Python calls that failed while a dependency was down 0 open listings

  discussion=https://news.ycombinator.com/item?id=49714038, title=Show HN: Replay the Python calls that failed while a dependency was down

• Repodify: Make Podcasts Out of Podcasts 0 open listings

  discussion=https://news.ycombinator.com/item?id=49743515, title=Show HN: Repodify: Make Podcasts Out of Podcasts

• Requirement5 0 open listings

  discussion=https://news.ycombinator.com/item?id=49699025, title=Show HN: Requirement5 – a digital card collection game, I used to collect Pokemon cards when I was a youth. I grew up and my Pokemon cards collected dust (and I learnt to be a software engineer). Then one day I stumbled across https://poke-holo.simey.me/ , which is an incredible piece of virtual Pokemon card work. Fuelled by this HTML, JS and CSS wizardry, I decided I'd build my own virtual collecting card game - which is call Requirement5. You discover cards, and collect our in game currency as you do so. If you want to save a card or create your own card, it costs currency. If you create a card and others save it, you get 70% of the currency back. A few friends and I have seeded it with our own creations (helped by the fact we have a CLI, so coding agents can do their bit). Our lore is about an interplanetary need for earth to express imagination: https://requirement5.com/about Discover cards: https://requirement5.com/ Create cards: https://requirement5.com/create Source: https://github.com/vilt9/requirement5

• Research-ready datasets from public data in mins 0 open listings

  discussion=https://news.ycombinator.com/item?id=49663006, title=Show HN: Research-ready datasets from public data in mins

• ResolveHQ 0 open listings

  discussion=https://news.ycombinator.com/item?id=49665864, title=Show HN: ResolveHQ – A Helpdesk Built on Cloudflare Workers, D1, R2 and Queues

• Respawn 0 open listings

  discussion=https://news.ycombinator.com/item?id=49721620, title=Show HN: Respawn – content-addressed filesystem snapshots and undo (Rust)

• Restarted 0 open listings

  discussion=https://news.ycombinator.com/item?id=49710968, title=Show HN: Restarted – a 2026 remake of the classic 2015 startup generator, The original startup website generator by Tiff Zhang and Mike Bradley landed on Hacker News in April 2015 and has been one of my favorite little novelties of that era ever since. It perfectly captures the saturated colors, cliché hero shots, gimmicky names, buzzword-heavy slogans, and proudly hirsute team photos of the time. A lot has changed since then, so I thought it would be fun to make a contemporary remake: https://restarted.io/ By default you get the minimalist aesthetic and clean-cut faces of 2026. The classic 2015 look is still available — just click the link at the bottom of the page or change the “z” parameter in the URL to the more familiar “s”. The universe of partner sites and competing startups is just as expansive as it ever was. The original site is entirely client-side and requires downloading all of the data tables locally. It leans on a mix of jQuery 1.11.2, Bootstrap 3.3.2, and Font Awesome 4.3.0, and if you view the source, it instantly gives away all of its secrets. For restarted.io I replaced all of that with a server-side renderer written in Go, so this time view-source tells you nothing. There are many Easter eggs in there — see how many you can find before I write them up. My original goal was to stay faithful to the 2015 appearance, and that turned out to be a technical adventure. The original's sine-based random number generator is... the worst, and different implementations of sine give different results. The eventual solution was to extract the exact sine function from Chrome’s V8 engine, as vendored C behind cgo and as a line-by-line Go port that keeps cgo optional, so the seeds and results line up the way they used to. Both are checked against V8’s own test cases. Then I discovered a bug in the original code that made half of its vocabulary unreachable — the first half of the verb table and the second half of the noun table, exactly complementary, so nothing about the output ever looked truncated. My goal then shifted from remaking the generator as it was in 2015 to remaking the site as the authors intended it to be in 2015. Over the years several people asked for their photos to be removed, so the remake instead draws from a broad pool of era-appropriate AI-generated profiles. A perceptual hash helps keep everyone looking distinct, and there’s a bit of extra care to make sure the Wang Fangs of the world don’t appear as Irish lasses. The hero image pool is much larger now, and all the old Rio de Janeiro shots have been retired, though you’ll still recognize plenty of the 2015 photos. Have fun poking around!

• Revera 0 open listings

  discussion=https://news.ycombinator.com/item?id=49665105, title=Show HN: Revera – local-first crypto API for 20 EVM chains, I built Revera to serve crypto prices locally with ultra-low latency. It supports 20 EVM chains and 3.6M+ tokens.

• Reverse-engineered a $800/mo indexing plan into Terraform and Flux 0 open listings

  discussion=https://news.ycombinator.com/item?id=49674768, title=Show HN: Reverse-engineered a $800/mo indexing plan into Terraform and Flux

• Reversing a ClickFix malware with a blockchain-based C2 0 open listings

  discussion=https://news.ycombinator.com/item?id=49655228, title=Show HN: Reversing a ClickFix malware with a blockchain-based C2

• Review agent changes locally before pushing 0 open listings

  discussion=https://news.ycombinator.com/item?id=49721156, title=Show HN: Review agent changes locally before pushing, I've often found that agents do stupid things when implementing requests. When I only see their changes with limited context in a scroll-by approval blob, it's easy to miss some of that stupidity. In those cases, I'll often only see it when I can review the full change in context in a PR. This tool adds a local review step, giving the human a final opportunity to be in the loop: your agent commits the changes, pushes to a local 'review' remote, and waits for feedback. The push itself opens a review via post-receive hook, and provides you with a localhost URL to review the changes. If you request changes via inline comments (like you would for a PR), the agent sees the comments and makes adjustments. It pushes again and the process repeats until you're satisfied. When you approve the change, your agent can take whatever step is next in your normal workflow - push to open an actual PR, merge locally, etc. These are lightweight instructions (~500 tokens) that should be customized for your workflow.

• Reviving my $15,000 browser arcade with Astra 0 open listings

  discussion=https://news.ycombinator.com/item?id=49651638, title=Show HN: Reviving my $15,000 browser arcade with Astra, Back in 2021, I put roughly $15,000 into a project called Anbu. The original idea was a decentralized cloud gaming platform, but I scaled it down to what I could afford with no cofounders... a browser arcade where people could play games with friends. I designed the platform myself in Figma, hand-drew the original art and logos, and commissioned development, pixel art, music, sound effects and some 3D models. That budget got us one game "Egg on the Run", along with game lobbies, friend invitations, chats, and a publishing platform. The plan was to attract players with the first game, have them bring friends, and eventually bring in other developers to publish their games. BUTTT it never got that far lol and the first game wasn’t compelling enough to attract much interest, and the cherry on top was the EC2 hosting cost around a fortune. AWS credits covered it initially, but once they ran out I couldn’t justify the expense so I shut it down never thought to see i'd see again. But then while working on my other startup Sky, I picked it up again as a weekend side project. I still had the original assets, and I wanted to see how much I could do with Astra. Now I am OBSESSED... EggVerse is now back online with an expanded arcade of 18 multiplayer games and more on the way! What’s been satisfying is returning to ideas I previously had to cut because I couldn’t afford to develop them. The original artwork and commissioned assets gave me a substantial starting point. Whether people enjoy it enough to return is still an open question. You can play at eggverse.co P.S. I’d suggest starting with Neon Nest (think mariokart) or grab a couple buddies/co-founders/employees and play Bag Egg (think imposter). I’d be interested to hear what you tried and whether you wanted to play another round. Looking forward to some feedback

• Revliu 0 open listings

  discussion=https://news.ycombinator.com/item?id=49625804, title=Show HN: Revliu – Multi-touch attribution from acquisition to revenue, I’m a founder and I just built RevliI’m a founder and I just built Revliu. After building several products, I always wanted to understand how my customers were actually finding me. Which campaign brought them in, which source they came from, and what their full journey looked like. There are already tools for this, but one problem I kept running into is that they often stop at first touch or last touch. That’s why I built Revliu around multi-touch attribution and the full customer journey. As soon as someone lands on your website, Revliu gives them an anonymous ID and starts tracking their journey. Once they sign up, that journey can be connected to the actual customer. From there, you can see if they’re on a trial, how often they come back, how many sessions they open, and eventually when they pay. For me, this also makes it easier to talk to that user and understand why they decided to move forward at that specific moment. Today, Revliu tracks acquisition sources, returning visitors, sessions, touchpoints, and revenue. Instead of just knowing that a campaign got a certain number of impressions, I want to know how much money it actually brought in. Impressions are not really what I care about. The app itself is pretty simple. The main page gives you an overview of revenue by source. Traffic shows visitors, their journey, and their sessions. Campaigns shows the acquisition funnel. For a SaaS, that can be visitor → lead → trial → paying customer. For an online store, it can simply be acquisition → purchase. Customers lets you look at each customer individually, see their touchpoints, where they came from, their journey, and when they bought. The technical side is simple in principle. It starts with an anonymous visitor who gets an ID. We keep track of their visits and touchpoints. When they sign up, that anonymous journey is connected to the customer. Stripe is connected as well, so when the customer pays, the revenue can be linked back to the same journey. The goal is basically to connect acquisition → visits → signup → customer → revenue. Right now, the app is still in its first phase. I haven’t built all the integrations I want yet. What I want to build next is something that does more than just show data. If someone leaves their email but never finishes signing up, comes back several times without buying, or shows real interest without converting, I want to use that information to try to recover that potential customer. The idea would be to reach them again by email or, depending on the integrations, through other channels like LinkedIn. That’s why I still see the app as limited today. I first want to test this attribution part and see if there’s real traction around the problem before building everything else. The part I’m really interested in for the future is revenue recovery. By combining it with AI, the goal would be to adapt to different customers, understand which ones are hesitating, and do more than just display data. I want the data to actually do something inside the product. And I’d really like to know how you’re handling this today. Coding has become much easier than it used to be, but the difference between a good product that sells and a good product nobody uses often comes down to marketing and distribution. I’d really like your feedback on Revliu, the way I’m approaching the problem, and what you would do differently.

• Rewic 0 open listings

  discussion=https://news.ycombinator.com/item?id=49713696, title=Show HN: Rewic – Turn Git commits into client-ready progress updates

• Ridge 0 open listings

  discussion=https://news.ycombinator.com/item?id=49647368, title=Show HN: Ridge - Connect coding agents to local, SSH, Docker, and S3 resources, I am working on an open source project called Ridge that gives coding agents a common interface to resources such as local projects, Docker containers, SSH machines, and S3 buckets. Agents can discover available resources, access data, copy between them, and execute commands where supported. It works through MCP, a CLI, or Python. I started building it because I wanted to use Codex on my laptop for kernel optimization experiments, with execution on my GPU box and models stored in S3. Before getting to the actual experiment, Codex needed to put together the file transfer and remote execution logic. You can save those scripts and leave instructions for the next session, but then you have a small infrastructure project to maintain alongside the experiment. The idea is to make these operations reusable across projects and environments. For example, an agent can copy a model from S3 to a GPU machine in one operation, without passing the file contents through its context. The initial providers implement common capability contracts that can be extended to other resources. Ridge also supports giving subagents access to selected resources and operations, tracking background jobs, and coordinating shared file updates. The agent harness still handles planning and launching workers. What do you think? Have you tried similar workflows with coding agents? I would be interested in hearing what you were trying to do, what got complicated, and what you ended up building or using to make it work.

• Ringee 0 open listings

  discussion=https://news.ycombinator.com/item?id=49701510, title=Show HN: Ringee – Open-source human and AI calling

• RistOS 0 open listings

  discussion=https://news.ycombinator.com/item?id=49625206, title=Show HN: RistOS – a de-Googled Pixel OS for your self-hosted LLM assistant

• Rockbox and Doom on iPod nano 3rd gen 0 open listings

  discussion=https://news.ycombinator.com/item?id=49729212, title=Show HN: Rockbox and Doom on iPod nano 3rd gen

• Rocky Surf 0 open listings

  discussion=https://news.ycombinator.com/item?id=49648776, title=Show HN: Rocky Surf – Open-source control plane for coding-agent cloud VMs, Rocky Surf started late last year after Claude Code crashed my laptop. The alpha version was designed as a commercial product, but I ditched the idea because it would have required significant capital. Basically, the business would have been a cloud reseller with a bit of software on top. I built the core for my own projects - simple stuff like bash commands to create, stop, start, and terminate EC2s. Added features along the way, like other clouds and making tool bundles so I could spin up a VM and automatically install Pi, OpenCode, etc, etc. Finally decided to turn it into an open source platform. It's pretty cool because a) it's easy to use; b) there's an MCP so servers can be managed outside of the UI; c) it's modular and the repo comes with skills to create new tool bundles or add other cloud providers. Claude Code, Codex, Cursor, etc all offer their own cloud-based infrastructure but they have to be controlled somewhere. This makes it so my controlling agent doesn't have to live on my laptop.

• Rokz: a browser port of a 1990 ASCII dungeon crawler 0 open listings

  discussion=https://news.ycombinator.com/item?id=49616059, title=Show HN: Rokz: a browser port of a 1990 ASCII dungeon crawler, For those younger than me, Kroz was Apogee's first game, a text-mode dungeon crawler from 1987. Apogee released the Turbo Pascal source under GPL in 2009. I ported the 1990 episode (Lost Adventures of Kroz) to the browser. No dependencies typescript. Some interesting tidbids I learned: The game's clock is its sound :-) Turbo Pascal's delay() blocks, and the sound effects are built out of delays, so a footstep is an 8 ms tone, 120 ms of silence, another 8 ms tone. That silence is the turn. You can't outrun Kroz by mashing keys because every move costs the length of its own beep. In the port a square-wave oscillator reproduces this (no samples etc), and playing a sound returns its duration whether or not it is audible. DOS Kroz asked at startup: "Slow or Fast PC" - and repeated sound chunks according to this. All the hidden stuff, easter eggs, and even most of bugs are reproduced. You can play it just like original. Oh, one thing - instead of ASCII letters I used tiles (from Dungeon Crawl Stone Soup (CC0)). I started with Lost Adventures because Apogee's source was most complete. I will continue with rest of the episodes. AI Disclose: I used Claude to help me extract levels from the Pascal source and to test it by taking over chrome session and measure miliseconds :D You can play it here: https://taler.pl/rokz/ source code here: https://github.com/htaler/rokz

• Roled 0 open listings

  discussion=https://news.ycombinator.com/item?id=49691404, title=Show HN: Roled – User and role management across multiple applications, Hi HN! I am building Roled ( https://roled.io ), an open-source platform for centralized user & role management across multiple applications. At several companies I've worked at, I noticed the same pattern: we built multiple products, and each needed users, roles, and permissions. Eventually we'd end up building some form of centralized backoffice or user management ourselves. I started Roled because I wondered whether that could simply be infrastructure that we don't have to rebuild. One design choice in Roled is that users aren't modeled around organizations. Instead, each project has its own users, roles, permissions, and authentication settings. This makes it possible to manage multiple independent applications from one place without requiring them to share the same user base. It provides ready-to-use authentication flows such as sign-up, sign-in, password reset, email verification, and user invitations. The user-management APIs can also be used directly from your backend if you want to build your own flows. Applications integrate with Roled using OAuth2. Roled is open source and can be self-hosted, or you can use the managed version at https://console.roled.io . The documentation is at https://docs.roled.io . I'd really appreciate it if you gave it a try, and I am looking forward to any feedback on your experience! I'm especially interested to know if any of you have a similar experience to mine. How did you solve it? Thanks!

• Rollstack Hiring 1 open listing

  employer=Rollstack, profile=https://www.ycombinator.com/companies/rollstack-2, article=https://thoughtbot.com/blog/48-companies-you-didn-t-know-were-using-rails-in-2026, source=https://www.ycombinator.com/, ruby_jobs=1, titles=AI Software Engineer

• Romy 0 open listings

  discussion=https://news.ycombinator.com/item?id=49642433, title=Show HN: Romy – finds first users for a product you've shipped

• Roomantic 0 open listings

  discussion=https://news.ycombinator.com/item?id=49611224, title=Show HN: Roomantic – See if that piece furniture fits before buying it, Roomantic is a project I started working on earlier this year when I learned that iPhone Pro models have a LiDAR scanner and almost no practical apps that take advantage of it.It uses Apple’s RoomPlan API to generate floor plans, and then allows you to paste in links to furniture and visualize if it would fit into your space.Every feature is free, a subscription option exists if you want unlimited furniture placements (since those actually cost me money).I tried my best to make the app beautiful, since features are now so cheap to build using AI, so any feedback would be appreciated.It primarily supports U.S. retailers, but if there is any international interest, I’d be happy to build that out in the future.Thanks!

• Rossino 0 open listings

  discussion=https://news.ycombinator.com/item?id=49736114, title=Show HN: Rossino – Pomodoro timer for Mac/iPhone/iPad/Watch, exact-second sync

• Round 1 of a vibe-coder ladder, judged by an LLM that only sees the UI 0 open listings

  discussion=https://news.ycombinator.com/item?id=49729103, title=Show HN: Round 1 of a vibe-coder ladder, judged by an LLM that only sees the UI

• Routi Bot 0 open listings

  discussion=https://news.ycombinator.com/item?id=49617611, title=Show HN: Routi Bot – AI bots with their own desktops on your Mac, been working on Routi Bot. an open source mac app inspired by Grok Bot. each bot has its own instructions, desktop and can be used with AI provider/model (openai astra, claude fable, deepseek, grok, etc). it also has a profile switcher so you can separate automation between your personal and work profiles.

• Ruby Hospital Kampala Hiring 1 open listing

  url=https://x.com/RubyhospitalUg/status/2098340015017263242, title=📢 We're Hiring! Ruby Hospital is looking for a Sales Executive – Entebbe to drive business development and patient referrals in the area. If you're proactive, w, matching_tags=Ruby, description=📢 We're Hiring! Ruby Hospital is looking for a Sales Executive – Entebbe to drive business development and patient referrals in the area. If you're proactive, well-connected, and based in Entebbe, apply before 18/09/2026 to hr@rubyhospitalk..., author=@RubyhospitalUg, author_name=Ruby Hospital Kampala

• Run coding agents on your own machine, steer them from anywhere 0 open listings

  discussion=https://news.ycombinator.com/item?id=49681398, title=Show HN: Run coding agents on your own machine, steer them from anywhere

• Run the birds sound detection on any Kobo eReader 0 open listings

  discussion=https://news.ycombinator.com/item?id=49743786, title=Show HN: Run the birds sound detection on any Kobo eReader, I saw a lot of people confused about how expensive the hardware parts ordering and assembly would be to run the awesome birds sound detection and rendering on e-ink frame posted this week here on HN. Here is how you can run that on a Kobo. A used can be bought for around US $50, much cheaper than buying the individual components, and it also doubles as an eBook reader. The Colour Kobo versions should be preferred over the BW ones. Also, add a nice case and it should look good.

• RunBoth, a behaviour diff for code an AI changed 0 open listings

  discussion=https://news.ycombinator.com/item?id=49686484, title=Show HN: RunBoth, a behaviour diff for code an AI changed

• Runyte 0 open listings

  discussion=https://news.ycombinator.com/item?id=49648804, title=Show HN: Runyte – a modal editor with integrated terminals and a file manager, Hi, I'm building Runyte, an open-source terminal workspace built around a modal

• Rust-split 0 open listings

  discussion=https://news.ycombinator.com/item?id=49671509, title=Show HN: Rust-split – save your tokens on large Rust source files, A common failure mode for AI SW development is that agents tend to build large "God" files. This is bad on your token burn. A) because the larger your source file the more churn the LLM takes to find and emit edits, which are more likely to fail, and B) the real kicker - the agent's process is to move one item at a time, which is an O(n^2) algorithm, but especially bad because it inevitably breaks and the code mushes up and sadness ensues (burning time and tokens). So, I have had lots of success with rust-split and a few skill notes. rust-split uses the AST (syn) to find the correct boundaries. Two passes: explode cuts the file into top-level items with their comments and #[attrs] attached - lossless, cat the chunks together and you get the original back byte for byte, so you can check it before trusting it - then split writes the module layout under a LOC ceiling. What's left for the LLM is basically fixing imports, which doesn't burn anywhere near as many tokens. It's not perfect with the transition but fixing small import issues LLMs can do efficiently. rust-split is released as a GitHub attested release or a crates.io install: https://github.com/owebeeone/rust-split cargo install rust-split

• S-Roll 0 open listings

  discussion=https://news.ycombinator.com/item?id=49743661, title=Show HN: S-Roll – local AI video understanding and clipping for Mac, Hi HN, I built S-Roll, a macOS app for turning long-form video into clips entirely on-device. Drop in podcasts, gameplay videos, or other long videos and get shorts or viral reels instantly. You can chat with the video to find a specific moment and turn it directly into a clip. S-Roll also does subject tracking, so the person or object you care about stays in frame when converting to vertical video. Everything runs locally on your Mac, including transcription and video understanding. S-Roll: https://saliency.dev/s-roll Discord: https://discord.gg/B6HnneePm Would love feedback :)

• SCIP MIP solver bindings for Go, ported from russcip 0 open listings

  discussion=https://news.ycombinator.com/item?id=49714495, title=Show HN: SCIP MIP solver bindings for Go, ported from russcip

• SF Title Alert 0 open listings

  discussion=https://news.ycombinator.com/item?id=49680001, title=Show HN: SF Title Alert

• SOS 0 open listings

  discussion=https://news.ycombinator.com/item?id=49657908, title=Show HN: SOS – Project state between coding agent sessions, I started working on SOS because noticed that during long-term projects, Codex often loses track of the current status and gets confused. I frequently had to explain what we had already done, which decisions were still valid, and what needed to be re-verified. Simple file-based governance didn't fully solve this problem. SOS stores a clearly defined "latest state" of the project within the repository. A current or new session retrieves this state via MCP. If something changes after a verification step, the previous result remains in the history but no longer influences the future development trajectory. This isn't automatic chat memory; the state must be explicitly saved and accepted. Missing information isn't filled in via guesswork. We have currently prepared an open-source alpha version focused on Codex; Claude support is planned for future releases. Linux is the primary supported OS. It has also been tested and works on macOS, while the Windows adapter is currently undergoing certification and will be available soon. SOS runs locally or on a server, if you prefer to work directly there. We do not use telemetry. I designed the system to be "agent-first," meaning you can delegate the setup to Codex simply by providing the repository link. The user just needs to provide confirmation at the appropriate steps. I’m interested to see how other users maintain agent state during long-term projects or when switching between agents. Do you find `AGENTS.md` or hand-off notes sufficient? Does governance solve the issue for you, and what challenges do you face? I am particularly interested in scenarios where an old instruction or a previously successful verification is no longer relevant to the current project.

• SQLite-Vector 0 open listings

  discussion=https://news.ycombinator.com/item?id=49659198, title=Show HN: SQLite-Vector – Vector search for SQLite, now Apache 2.0

Sign in to track