🛠️ All DevTools

[DevOps] Launch HN: Twill.ai (YC S25) – Delegate to cloud agents, get back PRs Hey HN, we&#x27;re Willy and Dan, co-founders of Twill.ai (<a href="https:&#x2F;&#x2F;twill.ai&#x2F;">https:&#x2F;&#x2F;twill.ai&#x2F;</a>). Twill runs coding CLIs like Claude Code and Codex in isolated cloud sandboxes. You hand it work through Slack, GitHub, Linear, our web app or CLI, and it comes back with a PR, a review, a diagnosis, or a follow-up question. It loops you in when it needs your input, so you stay in control.<p>Demo: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=oyfTMXVECbs" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=oyfTMXVECbs</a><p>Before Twill, building with Claude Code locally, we kept hitting three walls<p>1. Parallelization: two tasks that both touch your Docker config or the same infra files are painful to run locally at once, and manual port rebinding and separate build contexts don&#x27;t scale past a couple of tasks.<p>2. Persistence: close your laptop and the agent stops. We wanted to kick off a batch of tasks before bed and wake up to PRs.<p>3. Trust: giving an autonomous agent full access to your local filesystem and processes is a leap, and a sandbox per task felt safer to run unattended.<p>All three pointed to the same answer: move the agents to the cloud, give each task its own isolated environment.<p>So we built what we wanted. The first version was pure delegation: describe a task, get back a PR. Then multiplayer, so the whole team can talk to the same agent, each in their own thread. Then memory, so &quot;use the existing logger in lib&#x2F;log.ts, never console.log&quot; becomes a standing instruction on every future task. Then automation: crons for recurring work, event triggers for things like broken CI.<p>This space is crowded. AI labs ship their own coding products (Claude Code, Codex), local IDEs wrap models in your editor, and a wave of startups build custom cloud agents on bespoke harnesses. We take the following path: reuse the lab-native CLIs in cloud sandboxes. Labs will keep pouring RL into their own harnesses, so they only get better over time. That way, no vendor lock-in, and you can pick a different CLI per task or combine them.<p>When you give Twill a task, it spins up a dedicated sandbox, clones your repo, installs dependencies, and invokes the CLI you chose. Each task gets its own filesystem, ports, and process isolation. Secrets are injected at runtime through environment variables. After a task finishes, Twill snapshots the sandbox filesystem so the next run on the same repo starts warm with dependencies already installed. We chose this architecture because every time the labs ship an improvement to their coding harness, Twill picks up the improvement automatically.<p>We’re also open-sourcing agentbox-sdk, <a href="https:&#x2F;&#x2F;github.com&#x2F;TwillAI&#x2F;agentbox-sdk" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;TwillAI&#x2F;agentbox-sdk</a>, an SDK for running and interacting with agent CLIs across sandbox providers.<p>Here’s an example: a three-person team assigned Twill to a Linear backlog ticket about adding a CSV import feature to their Rails app. Twill cloned the repo, set up the dev environment, implemented the feature, ran the test suite, took screenshots and attached them to the PR. The PR needed one round of revision, which they requested through Github. For more complex tasks, Twill asks clarifying questions before writing code and records a browser session video (using Vercel&#x27;s Webreel) as proof of work.<p>Free tier: 10 credits per month (1 credit = $1 of AI compute at cost, no markup), no credit card. Paid plans start at $50&#x2F;month for 50 credits, with BYOK support on higher tiers. Free pro tier for open-source projects.<p>We’d love to hear how cloud coding agents fit into your workflow today, and if you try Twill, what worked, what broke, and what’s still missing.

[Other] Why I'm Building a Database Engine in C#

[Other] Show HN: Figma for Coding Agents Feels a bit like Figma, but for coding agents.<p>Instead of going back and forth with prompts, you give the agent a DESIGN.md that defines the design system up front, and it generally sticks to it when generating UI.<p>Google Stitch seems to be moving in this direction as a standard, so we put together a small collection of DESIGN.md files based on popular web sites.

[Other] Show HN: Zeroclawed: Secure Agent Gateway I’ve been cautiously (and nervously) playing with openclaw and a number of other claw and code agents for a while now, but trying out different ones was tricky so I wanted a simple way to switch out channel ownership… then I wanted more. Security is hard, and I wanted to make it easier. This is FAR from polished, and no claims that I’m a “security expert” but I tried to think and research a bit on different threat models (I think of 2 broad ones for agents, external adversaries and internal agentic failures) and try and offer best in class protection on both, while also not having any special opinion on what a good agent may look like today or in the future… this is just a gateway, and hopefully one that can work for nearly any agent now or in the future, but trying to come with batteries included for some of the more popular options today like openclaw, zeroclaw, claw-code, clause and opencode, not all there yet but contribution and critiques welcome.

[Other] CPU-Z and HWMonitor compromised <a href="https:&#x2F;&#x2F;xcancel.com&#x2F;vxunderground&#x2F;status&#x2F;2042483067655262461" rel="nofollow">https:&#x2F;&#x2F;xcancel.com&#x2F;vxunderground&#x2F;status&#x2F;2042483067655262461</a><p><a href="https:&#x2F;&#x2F;old.reddit.com&#x2F;r&#x2F;pcmasterrace&#x2F;comments&#x2F;1sh4e5l&#x2F;warning_hwmonitor_163_download_on_the_official&#x2F;" rel="nofollow">https:&#x2F;&#x2F;old.reddit.com&#x2F;r&#x2F;pcmasterrace&#x2F;comments&#x2F;1sh4e5l&#x2F;warni...</a><p><a href="https:&#x2F;&#x2F;www.bleepingcomputer.com&#x2F;news&#x2F;security&#x2F;supply-chain-attack-at-cpuid-pushes-malware-with-cpu-z-hwmonitor&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.bleepingcomputer.com&#x2F;news&#x2F;security&#x2F;supply-chain-...</a>

[Other] The acyclic e-graph: Cranelift's mid-end optimizer

[Code Quality] I shipped a transaction bug, so I built a linter

[Other] Show HN: A tool to manage a swarm of coding agents on Linux

[Other] Show HN: Run GUIs as Scripts Hi there, Zero Stars here.<p>I recently published some new work to Hokusai Pocket, which is a cross-platform binary made on top of raylib and MRuby that runs GUIs from ruby scripts.<p>License?<p>MIT!<p>How does it work?<p>The binary is available on the GitHub releases page: <a href="https:&#x2F;&#x2F;github.com&#x2F;skinnyjames&#x2F;hokusai-pocket&#x2F;releases&#x2F;tag&#x2F;0.6.1" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;skinnyjames&#x2F;hokusai-pocket&#x2F;releases&#x2F;tag&#x2F;0...</a><p>You can download the binary on x86 Windows, OSX, or Linux, and run your GUI application with<p>hokusai-pocket run:target=&quot;&lt;your_hokusai_app.rb&gt;&quot;<p>For a little bit of a hello world, I started a photoshop clone<p><a href="https:&#x2F;&#x2F;github.com&#x2F;skinnyjames&#x2F;hokusai_demo_paint" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;skinnyjames&#x2F;hokusai_demo_paint</a><p>Also a little game<p><a href="https:&#x2F;&#x2F;github.com&#x2F;skinnyjames&#x2F;pocket-squares" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;skinnyjames&#x2F;pocket-squares</a><p>Docs &#x2F; Help?<p>The docs are in progress, but the old docs for the CRuby version express some of the basic ideas around the project. <a href="https:&#x2F;&#x2F;hokusai.skinnyjames.net&#x2F;docs&#x2F;intro" rel="nofollow">https:&#x2F;&#x2F;hokusai.skinnyjames.net&#x2F;docs&#x2F;intro</a><p>(I&#x27;m also available to answer questions in between slinging pizza)<p>Deps?<p>Hokusai pocket currently uses<p>* libuv for offloading cpu intensive tasks to a worker pool to prevent blocking the UI thread, and I plan to integrate some libuv networking as well.<p>* raylib for backend graphics &#x2F; I&#x27;ve also built with SDL on arm64 to run applications on my pinephone<p>* NativeFileDialog for the lovely integration into filesystem.<p>* MRuby for running or embedding the scripts<p>* tree-sitter for the custom template grammar (Although templates can be built with ruby)<p>Anyway, I hope you get a chance to try it. If you make something cool AND have docker installed, you can also publish your work as single binary<p>`hokusai-pocket publish:target=&lt;your cool program.rb&gt;`<p>Would love feedback, apps, and help with documentation and more build targets.<p>urs truly,<p>@ ᴗ @

[Other] We've raised $17M to build what comes after Git

[Other] Show HN: SmolVM – open-source sandbox for coding and computer-use agents SmolVM is an open-source local sandbox for AI agents on macOS and Linux.<p>I started building it because agent workflows need more than isolated code execution. They need a reusable environment: write files in one step, come back later, snapshot state, pause&#x2F;resume, and increasingly interact with browsers or full desktop environments.<p>Right now SmolVM is a Python SDK and CLI focused on local developer experience.<p>Current features include: - local sandbox environments - macOS and Linux support - snapshotting - pause&#x2F;resume - persistent environments across turns<p>Install: ``` curl -sSL <a href="https:&#x2F;&#x2F;celesto.ai&#x2F;install.sh" rel="nofollow">https:&#x2F;&#x2F;celesto.ai&#x2F;install.sh</a> | bash smolvm ```<p>I’d love feedback from people building coding agents or computer-use agents. Interested in what feels missing, what feels clunky, and what you’d expect from a sandbox like this.

[Other] Show HN: Linear RNN/Reservoir hybrid generative model, one C file (no deps.) I just noticed it takes literally ~5 minutes to train millions parameters on slow CPU...but before you call Yudkowsky that &quot;it&#x27;s over&quot;, an important note: the main bottleneck is the corpus size, params are just &#x27;cleverness&#x27; but given limited info it&#x27;s powerless.<p>Anyway, here is the project:<p><a href="https:&#x2F;&#x2F;github.com&#x2F;bggb7781-collab&#x2F;lrnnsmdds&#x2F;tree&#x2F;main" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;bggb7781-collab&#x2F;lrnnsmdds&#x2F;tree&#x2F;main</a><p>couple of notes:<p>1. single C file, no dependencies. Below are literally all the &quot;dependencies&quot;, not even custom header (copy paste from the top of the single c file):<p>#define _POSIX_C_SOURCE 200809L<p>#include &lt;stdio.h&gt; #include &lt;stdlib.h&gt; #include &lt;string.h&gt; #include &lt;math.h&gt; #include &lt;time.h&gt; #include &lt;stdint.h&gt; #include &lt;stdbool.h&gt; #include &lt;float.h&gt; #include &lt;getopt.h&gt; #include &lt;errno.h&gt;<p>4136 lines of code in one file at the moment, that&#x27;s all.<p>2. easiest way to compile on Windows: download Cygwin (<a href="https:&#x2F;&#x2F;www.cygwin.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.cygwin.com&#x2F;</a>), then navigate to the directory where your lrnnsmdds.c file is and just run gcc on it with some optimizations, such as:<p>gcc -std=c17 -O3 -march=native --fast-math -o lrnn lrnnsmdds.c -lm<p>On Linux just run gcc, if for whatever reason you don&#x27;t have gcc on Linux do sudo &amp;&amp; apt-get install gcc --y ,or something...<p>On Apple: i&#x27;ve no idea or maybe just use vmware and install ubuntu and then run it.<p>Of course you can &#x27;git clone&#x27; and go to the dir, but again: it&#x27;s one file! copy it...<p>The repo has tiny toy corpus included where i&#x27;ve borrowed (hopefully it&#x27;s not plagiarism!) the name &quot;John Gordon&quot; from one of my favorite books &quot;Star Kings&quot;, by E. Hamilton. Just the first and last name are copied, the content is unique (well several poorly written sentences by myself...). Obviously it will overfit and result on copy-paste on such small corpus, the sole goal is to check if everything runs and not if it&#x27;s the A-G-I. You&#x27;d need your own 100kb+ if you want to generate unique meaningful text.<p>3. why&#x2F;what&#x2F;when&#x2F;how?<p>The github repo is self-explanatory i believe about features, uses and goals but in an attempt to summarize:<p>My main motivation was to create a fast alternative to transformers which works on CPU only, hence you see the bizarre&#x2F;not-easy task of doing this in C and not python and the lack of dependencies. In addition I was hoping it will also be clever alternative hence you see all those features more stacked than 90s BMW 850. The &#x27;reservoir&#x27; is the most novel feature though, it offers quick exact recall arguably different than RWKV 8 or the latest Mamba, in fact name of the architecture SMDDS comes from the first letters of the implemented features:<p>* S. SwiGLU in Channel Mixing (more coherence) * M. Multi-Scale Token Shift (larger context) * D. Data-Dependent Decay with Low-Rank (speed in large context) * D. Dynamic State Checkpointing (faster&#x2F;linear generation) * S. Slot-memory reservoir (perfect recall, transformers style).<p>If you face some issue just email me (easiest).<p>the good, the bad the ugly:<p>It is more or less working text-to-text novel alternative architecture, it&#x27;s not trying to imitate transformers nor LSTM, Mamba, RWKV though it shares many features with them - the bad is that it&#x27;s not blazing fast, if you&#x27;re armed with ryzen&#x2F;i7 16 cores or whatever and patience you can try training it on several small books via word tokenizer and low perplexity (under 1.2...) and see if it looks smarter&#x2F;faster. Since this is open source obviously the hope is to be improved: make it cuda-friendly, improve the features, port to python etc.<p>Depending on many factors I may try to push for v2 in July, August, September. My focus at the moment will be to test and scale since the features are many, it compiles with zero warnings on the 2 laptops i&#x27;ve tested(windows&#x2F;cygwin and ubuntu) and the speed is comparable to transformers. 10x!

[Testing] Hegel, a universal property-based testing protocol and family of PBT libraries

[Other] Instant 1.0, a backend for AI-coded apps

[Other] Show HN: Control your X/Twitter feed using a small on-device LLM We built a Chrome extension and iOS app that filters Twitter&#x27;s feed using Qwen3.5-4B for contextual matching. You describe what you don&#x27;t want in plain language—it removes posts that match semantically, not by keyword.<p>What surprised us was that because Twitter&#x27;s ranking algorithm adapts based on what you engage with, consistent filtering starts reshaping the recommendations over time. You&#x27;re implicitly signaling preferences to the algorithm. For some of us it &quot;healed&quot; our feed.<p>Currently running inference from our own servers with an experimental on-device option, and we&#x27;re working on fully on-device execution to remove that dependency. Latency is acceptable on most hardware but not great on older machines. No data collection; everything except the model call runs locally.<p>It doesn&#x27;t work perfectly (figurative language trips it up) but it&#x27;s meaningfully better than muting keywords and we use it ourselves every day.<p>Also promising how local &#x2F; open models can now start giving us more control over the algorithmic agents in our lives, because capability density is improving.

[Other] Research-Driven Agents: When an agent reads before it codes

[CLI Tool] Show HN: Mdpdf a 2k line C CLI to convert Markdown to tiny PDFs I have always searched for a simple tool which can convert a MD document to a well styled small PDF, just like a GitHub Readme.<p>This was inspired by TinyPDF <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46316968">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46316968</a><p>I also used this project to get familiar with agentic coding, which I had dreaded before.<p>mdpdf supports:<p>- using the included PDF fonts to generate tiny valid PDFs<p>- outputs as A4 or Letter depending on your locale<p>- plenty of common MD syntax: code blocks, inline code, lists, tables, and jpg and png images<p>That&#x27;s it. It covers probably most of the use cases and can help to simply convert a Markdown write-up to a PDF to share.<p>GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;schicho&#x2F;mdpdf" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;schicho&#x2F;mdpdf</a><p>A simple make call should build it for you.

[Build/Deploy] Show HN: I built a Cargo-like build tool for C/C++ I love C and C++, but setting up projects can sometimes be a pain.<p>Every time I wanted to start something new I&#x27;d spend the first hour writing CMakeLists.txt, figuring out find_package, copying boilerplate from my last project, and googling why my library isn&#x27;t linking. By the time the project was actually set up I&#x27;d lost all momentum.<p>So, I built Craft - a lightweight build and workflow tool for C and C++. Instead of writing CMake, your project configuration goes in a simple craft.toml:<p><pre><code> [project] name = &quot;my_app&quot; version = &quot;0.1.0&quot; language = &quot;c&quot; c_standard = 99 [build] type = &quot;executable&quot; </code></pre> Run craft build and Craft generates the CMakeLists.txt automatically and builds your project. Want to add dependencies? That&#x27;s just a simple command:<p><pre><code> craft add --git https:&#x2F;&#x2F;github.com&#x2F;raysan5&#x2F;raylib --links raylib craft add --path ..&#x2F;my_library craft add sfml </code></pre> Craft will clone the dependency, regenerate the CMake, and rebuild your project for you.<p>Other Craft features: craft init - adopt an existing C&#x2F;C++ project into Craft or initialize an empty directory. craft template - save any project structure as a template to be initialized later. craft gen - generate header and source files with starter boilerplate code. craft upgrade - keeps itself up to date.<p>CMakeLists.extra.cmake for anything that Craft does not yet handle.<p>Cross platform - macOS, Linux, Windows.<p>It is still early (I just got it to v1.0.0) but I am excited to be able to share it and keep improving it.<p>Would love feedback. Please also feel free to make pull requests if you want to help with development!

[Other] The Vercel plugin on Claude Code wants to read your prompts

[API/SDK] Show HN: Zoneless – Open-source Stripe Connect clone with $0.002 fees using USDC Hi HN,<p>I&#x27;m Ben &#x2F; Tiny Projects (I once posted here about buying 300 emoji domains from Kazakhstan…).<p>For the past 3 years I&#x27;ve been solo bootstrapping PromptBase, an AI marketplace with 450k+ users. At the peak, I was burning $9,400&#x2F;month in opaque Stripe Connect fees for seller payouts, so I built Zoneless to replace it:<p>- GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;zonelessdev&#x2F;zoneless" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;zonelessdev&#x2F;zoneless</a><p>- Website: <a href="https:&#x2F;&#x2F;zoneless.com" rel="nofollow">https:&#x2F;&#x2F;zoneless.com</a><p>Zoneless is a free, open-source (Apache 2.0) drop-in replacement for the payout part of Stripe Connect. It allows you to pay marketplace sellers globally with stablecoins (USDC) using an identical API to Stripe and at near-zero fees.<p>I&#x27;ve been dogfooding Zoneless on PromptBase for 3 months with some good results:<p>- 2,200+ sellers onboarded<p>- 1,400+ payouts completed<p>- Monthly payout fees reduced to just a few dollars<p>- 73% of sellers, when given the choice at onboarding, actively picked Zoneless over Stripe Connect<p>A massive part of running a marketplace is paying sellers. While Stripe Connect is a great product, it has big pain points:<p>- Expensive + complex fees: <i>$2&#x2F;mo per active account, 0.25% + $0.25 domestic payout fee, $1.50 international payout fee, 0.25–1.25% cross-border fee, 0.50–1% FX fee.</i> It costs &gt;$2 to move $1.<p>- Limited reach: Only supports around 47 countries.<p>- Slow payouts: Takes 2-7 days to settle.<p>- Platform risk: A massive single point of failure if your account gets randomly flagged.<p>Zoneless is designed to solve all this:<p>- Payouts cost ~$0.002 on Solana<p>- Global: 220+ countries&#x2F;regions<p>- Instant payouts, 24&#x2F;7.<p>- Self-hostable and open-source<p>The API&#x2F;SDK is identical to Stripe (same webhook events, same object shapes, etc.). If you know Stripe, you already know how to use Zoneless. There’s also an Express Dashboard for sellers to onboard and track their earnings.<p>I&#x27;ve been able to remove annoying things on PromptBase like forcing sellers to accrue a $30 minimum balance before a payout just to keep our costs down. I can now also onboard sellers from more countries, which has helped spread the word and grow the buyer side too.<p>A big worry was that non-crypto users would be confused or hate getting paid in USDC, but they actually don’t mind at all, they just care about being paid faster. If they want to convert to their local currency, they simply use an exchange like Coinbase.<p>Zoneless is self-custodial, meaning you create and own your wallet, and the code never touches funds. You can also easily plug in providers for KYC&#x2F;AML.<p>I appreciate that anything related to crypto is like Marmite (pretty polarizing); I’m a no-coiner and have never dabbled in NFTs, but I do think stablecoins are different: they’re just boring tech to move money around cheaply.<p>I&#x27;d love to hear your thoughts, feedback, or questions - especially if you&#x27;ve dealt with Stripe Connect &#x2F; payouts &#x2F; marketplaces before.<p>- GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;zonelessdev&#x2F;zoneless" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;zonelessdev&#x2F;zoneless</a><p>- Website: <a href="https:&#x2F;&#x2F;zoneless.com" rel="nofollow">https:&#x2F;&#x2F;zoneless.com</a><p>- Docs: <a href="https:&#x2F;&#x2F;zoneless.com&#x2F;docs" rel="nofollow">https:&#x2F;&#x2F;zoneless.com&#x2F;docs</a>