🛠️ All DevTools

Sub2API-CRS2 一站式开源中转服务，让 Claude、Openai 、Gemini、Antigravity订阅统一接入，支持拼车共享，更高效分摊成本，原生工具无缝使用。

💖🧸 Self hosted, you-owned Grok Companion, a container of souls of waifu, cyber livings to bring them into our worlds, wishing to achieve Neuro-sama's altitude. Capable of realtime voice chat, Minecraft, Factorio playing. Web / macOS / Windows supported.

Show HN: Xmloxide – an agent-made Rust replacement for libxml2 Recently several AI labs have published experiments where they tried to get AI coding agents to complete large software projects.<p>- Cursor attempted to make a browser from scratch: <a href="https:&#x2F;&#x2F;cursor.com&#x2F;blog&#x2F;scaling-agents" rel="nofollow">https:&#x2F;&#x2F;cursor.com&#x2F;blog&#x2F;scaling-agents</a><p>- Anthropic attempted to make a C Compiler: <a href="https:&#x2F;&#x2F;www.anthropic.com&#x2F;engineering&#x2F;building-c-compiler" rel="nofollow">https:&#x2F;&#x2F;www.anthropic.com&#x2F;engineering&#x2F;building-c-compiler</a><p>I have been wondering if there are software packages that can be easily reproduced by taking the available test suites and tasking agents to work on projects until the existing test suites pass.<p>After playing with this concept by having Claude Code reproduce redis and sqlite, I began looking for software packages where an agent-made reproduction might actually be useful.<p>I found libxml2, a widely used, open-source C language library designed for parsing, creating, and manipulating XML and HTML documents. Three months ago it became unmaintained with the update, &quot;This project is unmaintained and has [known security issues](<a href="https:&#x2F;&#x2F;gitlab.gnome.org&#x2F;GNOME&#x2F;libxml2&#x2F;-&#x2F;issues&#x2F;346" rel="nofollow">https:&#x2F;&#x2F;gitlab.gnome.org&#x2F;GNOME&#x2F;libxml2&#x2F;-&#x2F;issues&#x2F;346</a>). It is foolish to use this software to process untrusted data.&quot;.<p>With a few days of work, I was able to create xmloxide, a memory safe rust replacement for libxml2 which passes the compatibility suite as well as the W3C XML Conformance Test Suite. Performance is similar on most parsing operations and better on serialization. It comes with a C API so that it can be a replacement for existing uses of libxml2.<p>- crates.io: <a href="https:&#x2F;&#x2F;crates.io&#x2F;crates&#x2F;xmloxide" rel="nofollow">https:&#x2F;&#x2F;crates.io&#x2F;crates&#x2F;xmloxide</a><p>- GitHub release: <a href="https:&#x2F;&#x2F;github.com&#x2F;jonwiggins&#x2F;xmloxide&#x2F;releases&#x2F;tag&#x2F;v0.1.0" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;jonwiggins&#x2F;xmloxide&#x2F;releases&#x2F;tag&#x2F;v0.1.0</a><p>While I don&#x27;t expect people to cut over to this new and unproven package, I do think there is something interesting to think about here in how coding agents like Claude Code can quickly iterate given a test suite. It&#x27;s possible the legacy code problem that COBOL and other systems present will go away as rewrites become easier. The problem of ongoing maintenance to fix CVEs and update to later package versions becomes a larger percentage of software package management work.

Show HN: I built a tool to translate and declutter articles for my immigrant mom Hello HN,<p>I built DuLink to solve a personal problem: my primary language is English, but my mother’s primary language is Mandarin Chinese. I often find articles on health or current events that I want to share with her, but the friction of copy&#x2F;pasting and Google Translate meant she rarely read them.<p>What it does:<p>DuLink takes an article URL, extracts the core content to strip out clutter, and then generates a static, translated reading view. It preserves the semantic HTML structure (headers, lists) but removes the original site&#x27;s JS and ad tracking.<p>The &quot;Why&quot;:<p>Existing tools either translate the entire messy page (including nav menus&#x2F;footers) or require the recipient to install an extension. I wanted something truly simple. A &quot;fire and forget&quot; link that removed every obstacle. No setup, no extra tools, just open and understand. I also added audio playback because reading dense text on a phone can be tiring, especially for aging eyes. Sometimes listening is simply easier.<p>I built this for a personal use case, but I&#x27;m sharing it here for anyone who wants to share something meaningful across languages. If it helps you bridge that gap, I&#x27;d love to hear about it.<p>Any feedback is welcome.<p><a href="https:&#x2F;&#x2F;dulink.click&#x2F;" rel="nofollow">https:&#x2F;&#x2F;dulink.click&#x2F;</a>

Building a Minimal Transformer for 10-digit Addition

Show HN: Tomoshibi – A writing app where your words fade by firelight I spent ten years trying to write a novel. Every time I sat down, I&#x27;d write a sentence, decide it wasn&#x27;t good enough, and rewrite it.<p>The problem wasn&#x27;t discipline — it was that I could always see what I&#x27;d written and go back to change it.<p>I tried other approaches. Apps that delete your words when you stop typing — they fight fear with fear. That just made me panic. I wanted the opposite: not punishment, but permission.<p>&quot;Tomoshibi&quot; is Japanese for a small light in the dark — just enough to see what&#x27;s in front of you.<p>You write on a dark screen. Older lines fade, but not when you hit return. They fade when you start writing again. If you pause, they wait. You can edit the current line and one line back — enough to fix a typo, not enough to spiral. The one-line-back rule also catches my own practical issue: Japanese IME often fires an accidental newline on kanji confirmation.<p>Everything is saved. There&#x27;s a separate reader view for going back through what you&#x27;ve written. Tomoshibi is for writing over months, not just one session. When you come back, your last sentence appears as an epigraph — as if it always belonged there.<p>No account, no server, no build step. Your writing stays in your browser&#x27;s local storage — export anytime as .txt. Vanilla HTML&#x2F;CSS&#x2F;ES modules.<p>Try it in your browser. A native Mac app (built with Tauri) with file system integration is coming to the store.<p>I&#x27;ve been writing on it for two months.<p><a href="https:&#x2F;&#x2F;tomoshibi.in-hakumei.com&#x2F;app&#x2F;" rel="nofollow">https:&#x2F;&#x2F;tomoshibi.in-hakumei.com&#x2F;app&#x2F;</a>

[Other] Show HN: Mowgli – Figma for the agent era, with Claude Code and design export Hi HN! We&#x27;re excited to announce the public beta of Mowgli, a spec-backed, AI-native design canvas for scoping and ideating on products.<p>The productivity gains unleashed by coding agents have made everything else an unexpected bottleneck. In an effort to make the most out of this new paradigm, we ceded a lot of ground in product thinking, thoughtful UX, and design excellence. In other words, the pace of tooling for deciding <i>what</i> to build has not kept up.<p>Mowgli is inspired by, in equal parts, Figma and Claude&#x27;s plan mode. It evolves a detailed specification and designs for every screen and state of the product on an infinite canvas. Owing to this UI, it can quickly mock up new features and flows, redesign existing ones, and show you variations side by side. LLMs are amazing at helping you explore vast solution spaces, but most current tooling focuses on getting narrowly perfect output based on a well-defined spec. We try to bridge that gap with Mowgli.<p>When you&#x27;re ready to build, Mowgli offers a .zip export with a SPEC.md and unopinionated design .tsx files for your screens, in a perfect format for coding agents. Or alternatively, full, pixel-perfect export of all screens to Figma.<p>In this early stage, we support two entrypoints: (1) building a product from scratch through a guided experience, or (2) importing an existing product from Figma. We have a powerful, almost pixel-perfect Figma to code + spec pipeline that works on files of any size - from 0 to 300+ frames.<p>We&#x27;re working hard on other ways to get your existing products into Mowgli (and would appreciate hearing about what you would like!)<p>- Timelapse of making a functional second brain app in Mowgli + Claude Code: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=HeOoy8WDmMA" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=HeOoy8WDmMA</a><p>- Sample project designed and specced entirely by Mowgli (demo, no login needed): <a href="https:&#x2F;&#x2F;app.mowgli.ai&#x2F;projects&#x2F;cmluzdfa0000v01p91l5r61e3?theme=dark&amp;demo=true&amp;showchat=true&amp;dismisschat=true" rel="nofollow">https:&#x2F;&#x2F;app.mowgli.ai&#x2F;projects&#x2F;cmluzdfa0000v01p91l5r61e3?the...</a><p>- Sample Figma import of the whole Posthog product, ready for iteration: <a href="https:&#x2F;&#x2F;app.mowgli.ai&#x2F;projects&#x2F;cmkl0zqng000101lo8yn2gqvd?demo=true&amp;theme=dark" rel="nofollow">https:&#x2F;&#x2F;app.mowgli.ai&#x2F;projects&#x2F;cmkl0zqng000101lo8yn2gqvd?dem...</a> (thanks PostHog for being radically open with public Figma files!)

[Other] Obsidian Sync now has a headless client

[Database] Show HN: SQLite for Rivet Actors – one database per agent, tenant, or document Hey HN! We posted Rivet Actors here previously [1] as an open-source alternative to Cloudflare Durable Objects.<p>Today we&#x27;ve released SQLite storage for actors (Apache 2.0).<p>Every actor gets its own SQLite database. This means you can have millions of independent databases: one for each agent, tenant, user, or document.<p>Useful for:<p>- AI agents: per-agent DB for message history, state, embeddings<p>- Multi-tenant SaaS: real per-tenant isolation, no RLS hacks<p>- Collaborative documents: each document gets its own database with built-in multiplayer<p>- Per-user databases: isolated, scales horizontally, runs at the edge<p>The idea of splitting data per entity isn&#x27;t new: Cassandra and DynamoDB use partition keys to scale horizontally, but you&#x27;re stuck with rigid schemas (&quot;single-table design&quot; [3]), limited queries, and painful migrations. SQLite per entity gives you the same scalability without those tradeoffs [2].<p>How this compares:<p>- Cloudflare Durable Objects &amp; Agents: most similar to Rivet Actors with colocated SQLite and compute, but closed-source and vendor-locked<p>- Turso Cloud: Great platform, but closed-source + diff use case. Clients query over the network, so reads are slow or stale. Rivet&#x27;s single-writer actor model keeps reads local and fresh.<p>- D1, Turso (the DB), Litestream, rqlite, LiteFS: great tools for running a single SQLite database with replication. Rivet is for running lots of isolated databases.<p>Under the hood, SQLite runs in-process with each actor. A custom VFS persists writes to HA storage (FoundationDB or Postgres).<p>Rivet Actors also provide realtime (WebSockets), React integration (useActor), horizontal scalability, and actors that sleep when idle.<p>GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;rivet-dev&#x2F;rivet" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;rivet-dev&#x2F;rivet</a><p>Docs: <a href="https:&#x2F;&#x2F;www.rivet.dev&#x2F;docs&#x2F;actors&#x2F;sqlite&#x2F;">https:&#x2F;&#x2F;www.rivet.dev&#x2F;docs&#x2F;actors&#x2F;sqlite&#x2F;</a><p>[1] <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=42472519">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=42472519</a><p>[2] <a href="https:&#x2F;&#x2F;rivet.dev&#x2F;blog&#x2F;2025-02-16-sqlite-on-the-server-is-misunderstood&#x2F;">https:&#x2F;&#x2F;rivet.dev&#x2F;blog&#x2F;2025-02-16-sqlite-on-the-server-is-mi...</a><p>[3] <a href="https:&#x2F;&#x2F;www.alexdebrie.com&#x2F;posts&#x2F;dynamodb-single-table&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.alexdebrie.com&#x2F;posts&#x2F;dynamodb-single-table&#x2F;</a>

[Other] Official data.gouv.fr Model Context Protocol (MCP) server that allows AI chatbots to search, explore, and analyze datasets from the French national Open Data platform, directly through conversation.

[Other] Show HN: Rust-powered document chunker for RAG – 40x faster, O(1) memory I built a document chunking library for RAG pipelines with a Rust core and Python bindings.<p>The problem: LangChain&#x27;s chunker is pure Python and becomes a bottleneck at scale — slow and memory-hungry on large document sets.<p>What Krira Chunker does differently: - Rust-native processing — 40x faster than LangChain&#x27;s implementation - O(1) space complexity — memory stays flat regardless of document size - Drop-in Python API — works with any existing RAG pipeline - Production-ready — 17 versions shipped, 315+ installs<p>pip install krira-augment<p>Would love brutal feedback from anyone building RAG systems — what chunking problems are you running into that this doesn&#x27;t solve yet?

Show HN: Decided to play god this morning, so I built an agent civilisation at a pub in london, 2 weeks ago - I asked myself, if you spawned agents into a world with blank neural networks and zero knowledge of human existence — no language, no economy, no social templates — what would they evolve on their own?<p>would they develop language? would they reproduce? would they evolve as energy dependent systems? what would they even talk about?<p>so i decided to make myself a god, and built WERLD - an open-ended artificial life sim, where the agent&#x27;s evolve their own neural architecture.<p>Werld drops 30 agents onto a graph with NEAT neural networks that evolve their own topology, 64 sensory channels, continuous motor effectors, and 29 heritable genome traits. communication bandwidth, memory decay, aggression vs cooperation — all evolvable. No hardcoded behaviours, no reward functions. - they could evolve in any direction.<p>Pure Python, stdlib only — brains evolve through survival and reproduction, not backprop. There&#x27;s a Next.js dashboard (&quot;Werld Observatory&quot;) that gives you a live-view: population dynamics, brain complexity, species trajectories, a narrative story generator, live world map.<p>thought this would be more fun as an open-source project!<p>can&#x27;t wait to see where this could evolve - i&#x27;ll be in the comments and on the repo.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;nocodemf&#x2F;werld" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;nocodemf&#x2F;werld</a>

OpenAI fires an employee for prediction market insider trading

[Other] Show HN: Now I Get It – Translate scientific papers into interactive webpages Understanding scientific articles can be tough, even in your own field. Trying to comprehend articles from others? Good luck.<p>Enter, Now I Get It!<p>I made this app for curious people. Simply upload an article and after a few minutes you&#x27;ll have an interactive web page showcasing the highlights. Generated pages are stored in the cloud and can be viewed from a gallery.<p>Now I Get It! uses the best LLMs out there, which means the app will improve as AI improves.<p>Free for now - it&#x27;s capped at 20 articles per day so I don&#x27;t burn cash.<p>A few things I (and maybe you will) find interesting:<p>* This is a pure convenience app. I could just as well use a saved prompt in Claude, but sometimes it&#x27;s nice to have a niche-focused app. It&#x27;s just cognitively easier, IMO.<p>* The app was built for myself and colleagues in various scientific fields. It can take an hour or more to read a detailed paper so this is like an on-ramp.<p>* The app is a place for me to experiment with using LLMs to translate scientific articles into software. The space is pregnant with possibilities.<p>* Everything in the app is the result of agentic engineering, e.g. plans, specs, tasks, execution loops. I swear by Beads (<a href="https:&#x2F;&#x2F;github.com&#x2F;steveyegge&#x2F;beads" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;steveyegge&#x2F;beads</a>) by Yegge and also make heavy use of Beads Viewer (<a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46314423">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46314423</a>) and Destructive Command Guard (<a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46835674">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46835674</a>) by Jeffrey Emanuel.<p>* I&#x27;m an AWS fan and have been impressed by Opus&#x27; ability to write good CFN. It still needs a bunch of guidance around distributed architecture but way better than last year.

Show HN: Reclaim Flowers – A 2D physics-based "Digital Altar" protocol

[Other] Show HN: Gitcredits – movie-style end credits for any Git repo in your terminal

[Other] PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）

[Other] IDE for the AI Agents Era - Run an army of Claude Code, Codex, etc. on your machine

[Other] Show HN: SplatHash – A lightweight alternative to BlurHash and ThumbHash Hi HN,<p>I built SplatHash. It&#x27;s a lightweight image placeholder generator I wrote to be a simpler, faster alternative to BlurHash and ThumbHash.<p>Repo: <a href="https:&#x2F;&#x2F;github.com&#x2F;junevm&#x2F;splathash" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;junevm&#x2F;splathash</a>

[Other] Stop Burning Your Context Window – How We Cut MCP Output by 98% in Claude Code