🛠️ Hacker News Tools

[Other] Show HN: Local Browser AI Using the new Prompt API to build a free open source privacy-first chat extension that runs on Chrome or Edge on Linux, Mac, Windows and ChromeOS

[Other] Show HN: Run PyTorch locally with a remote GPU backend I integrated a remote GPU execution backend into PyTorch through the same system that custom hardware accelerators get integrated into PyTorch. You can create a remote machine and obtain its CUDA device whenever you want to create or move tensors onto the remote GPU.<p><pre><code> machine = mycelya_torch.RemoteMachine(&quot;modal&quot;, &quot;A100&quot;) cuda_device = machine.device(&quot;cuda&quot;) x = torch.randn(1000, 1000, device=cuda_device) y = torch.randn(1000, 1000).to(cuda_device) </code></pre> I made it reasonably performant by having most operations dispatch asynchronously whenever possible. For cases where slow performance is unavoidable such as uploading many GB of weights onto the GPU, I added a decorator that can be applied to functions to turn it into a remotely executed function. For the most part, the function should behave the same with or without the decorator; the main difference is whether the function code is executed locally or remotely.<p><pre><code> import mycelya_torch from transformers import AutoModelForCausalLM, AutoTokenizer @mycelya_torch.remote def load_model(model_name: str): tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForCausalLM.from_pretrained( model_name, torch_dtype=&quot;auto&quot;, device_map=&quot;auto&quot; ) return model, tokenizer </code></pre> You can only use it with Modal as the cloud provider right now, and it&#x27;s free to use with their free monthly credits. I appreciate any feedback and bug reports :)

[Other] Show HN: Hyprvoice – Voice-Powered Typing for Wayland/Hyprland (No X11 Hacks) I built Hyprvoice, a small tool that lets you type with your voice on Wayland — no X11, no kludgy bridges. Press a key, talk, and your words appear wherever your cursor is.<p>It’s written in Go, uses PipeWire for audio, and talks directly to the compositor. The workflow is simple:<p>press key → start recording<p>press again → stop and inject text<p>Notifications show recording&#x2F;transcribing states, and it works with multiple backends — currently OpenAI Whisper, with whisper.cpp (local&#x2F;offline) in progress.<p>Everything runs through a lightweight daemon with IPC over a unix socket. Text injection uses wl-clipboard + wtype with clipboard restore fallback.<p>Install (Arch&#x2F;AUR):<p>yay -S hyprvoice-bin systemctl --user enable --now hyprvoice.service<p>Then add something like this to Hyprland:<p>bind = SUPER, R, exec, hyprvoice toggle<p>Repo: <a href="https:&#x2F;&#x2F;github.com&#x2F;leonardotrapani&#x2F;hyprvoice" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;leonardotrapani&#x2F;hyprvoice</a><p>It’s beta but fully usable. I’d love feedback, especially from people running Wayland full-time or maintaining compositors.

[Other] Show HN: AI toy I worked on is in stores Alt link: <a href="https:&#x2F;&#x2F;mrchristmas.com&#x2F;products&#x2F;santas-magical-telephone" rel="nofollow">https:&#x2F;&#x2F;mrchristmas.com&#x2F;products&#x2F;santas-magical-telephone</a><p>Video demo: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=0z7QJxZWFQg" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=0z7QJxZWFQg</a><p>The first time I talked with AI santa and it responded with a joke I was HOOKED. The fun&#x2F;nonsense doesn&#x27;t click until you try it yourself. What&#x27;s even more exciting is you can build it yourself:<p>libpeer: <a href="https:&#x2F;&#x2F;github.com&#x2F;sepfy&#x2F;libpeer" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;sepfy&#x2F;libpeer</a><p>pion: <a href="https:&#x2F;&#x2F;github.com&#x2F;pion&#x2F;webrtc" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;pion&#x2F;webrtc</a><p>Then go do all your fun logic in your Pion server. Connect to any Voice AI provider, or roll your own via Open Source. Anything is possible.<p>If you have questions or hit any roadblocks I would love to help you. I have lots of hardware snippets on my GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;sean-der" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;sean-der</a>.

[Other] Show HN: Vello's high-performance 2D GPU engine to .NET VelloSharp v0.5.0-alpha.1 is now available, delivering Vello’s high-performance 2D GPU engine to .NET atop the wgpu 3D backend. Integrates cleanly with Avalonia, WPF, and WinForms to modernize rendering pipelines without disruptive rewrites.

[Other] A Guide for WireGuard VPN Setup with Pi-Hole Adblock and Unbound DNS

[Other] Show HN: Rebuilt Bible search app to run 100% client-side with Transformers.js I rebuilt Biblos, my semantic Bible search app, to run entirely in your browser. No more server costs. The main challenge was fitting an accurate text embedding model into browsers. Last year&#x27;s version cost $20 monthly to host. The new version runs free on Vercel and searches 31k Bible verses without sending data to any server.<p>I pre-compute embeddings for all 31,000 Bible verses offline using BGE-large-en-v1.5. Each verse becomes a 1024-dimensional vector stored as JSON, compressed into ZIP files by book. When you visit the site, your browser downloads Transformers.js and the BGE model.<p>The model generates an embedding for your query, then finds similar verses using cosine similarity against the cached embeddings. Everything happens in your browser.<p>The app is designed to search Bible verses by semantic meaning, not just keywords. There&#x27;s a separate tab to read daily passages with church fathers commentary from 15+ sources, based on a read the bible in a year plan.<p>Free to use, no sign-in features or anything like that. Try it out biblos.app. I&#x27;d appreciate feedback on the search experience, especially on mobile devices! Happy to discuss the technical details or help anyone building similar browser-based ML applications.

[Other] Tangled, a Git collaboration platform built on atproto

[Other] Show HN: Modeling the Human Body in Rust So I Can Cmd+Click Through It I started this trying to understand two things: why my Asian friends turn red after drinking, and why several friends all seemed to have migraine clusters.<p>I was reading medical papers and textbooks, but kept getting lost jumping between topics. I thought: what if I could just Cmd+Click through this like code? What if &quot;ALDH2 gene&quot; was actually clickable, and took me to the variant, the phenotype, the population frequencies?<p>So I started modeling human biology in Rust with my Ralph agent (Claude in a loop, ty ghuntley). Turns out the type system is perfect for this. Every biological entity is strongly-typed with relationships enforced at compile time.<p>After 1 day of agent coding: - 277 Rust files, ~95k lines of code - 1,561 tests passing - 13 complete organ systems - Genetics with ancestry-specific variants - Clinical pathology models<p>Try it:<p>git clone <a href="https:&#x2F;&#x2F;github.com&#x2F;lantos1618&#x2F;open_human_ontology" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;lantos1618&#x2F;open_human_ontology</a> cd open_human_ontology cargo run --example ide_navigation_demo<p>Then open `examples&#x2F;ide_navigation_demo.rs` and Cmd+Click through:<p>Understanding Asian flush:<p>AsianGeneticVariantsCatalog::get_metabolic_variants()<p>&#x2F;&#x2F; Click through to:<p>&#x2F;&#x2F; → ALDH2 gene on chromosome 12q24.12<p>&#x2F;&#x2F; → rs671 variant (Glu504Lys)<p>&#x2F;&#x2F; → 40% frequency in Japanese population<p>&#x2F;&#x2F; → Alcohol flush reaction<p>&#x2F;&#x2F; → 10x esophageal cancer risk with alcohol<p>&#x2F;&#x2F; → Acetaldehyde metabolism pathway<p>Understanding migraines: Migraine { subtype: WithAura, triggers: [Stress, LackOfSleep, HormonalChanges], genetic_variants: [&quot;rs2075968&quot;, &quot;rs1835740&quot;], ... }<p>&#x2F;&#x2F; Click through to:<p>&#x2F;&#x2F; → 17 migraine trigger types<p>&#x2F;&#x2F; → 12 aura symptom types<p>&#x2F;&#x2F; → Genetic risk factors<p>&#x2F;&#x2F; → Why clusters happen (HormonalChanges → Menstruation)<p>Now I can actually <i>navigate</i> the connections instead of flipping through PDFs. Heart → CoronaryArtery → Plaque. VisualCortex → 200M neurons → NeuralConnection pathways. It&#x27;s like Wikipedia but type-checked and with jump-to-definition.<p>This isn&#x27;t production medical software - it&#x27;s a learning tool. But it&#x27;s way more useful than textbooks for understanding how biological systems connect.<p>The agent keeps expanding it. Sometimes it OOMs but that&#x27;s part of the fun.<p>Tech: Rust, nalgebra, serde, rayon, proptest<p>I am not a dr or medical professional this is for my education you can commit to it if you want to or review and open some PR&#x27;s if you find wrong information or want to add references.

[Other] Show HN: Gitcasso – Syntax Highlighting and Draft Recovery for GitHub Comments I built a browser extension called Gitcasso which:<p>- Adds markdown syntax highlighting to GitHub textareas<p>- Lists every open PR&#x2F;issue tab and any drafts<p>- (Optional, unimplemented) autosaves your comment drafts so you don’t lose work<p>I made it because I was impressed by <a href="https:&#x2F;&#x2F;overtype.dev&#x2F;" rel="nofollow">https:&#x2F;&#x2F;overtype.dev&#x2F;</a> (a markdown textarea syntax highlighter) which went big on here on HN a few weeks ago, and it seemed like a perfect fit for a GitHub browser extension. Keeping up with changes on upstream GitHub would normally be a pain, but with with Playwright and Claude Code it seemed possible for it to be nearly automatic, which has turned out to be mostly true!<p>This was the first time where I built a tool, gave the tool to AI, and then AI used the tool to make the thing I hoped it would be able to make. I&#x27;m pretty sold on the general technique...<p>GitHub repo (Apache2-licensed, open source): <a href="https:&#x2F;&#x2F;github.com&#x2F;diffplug&#x2F;gitcasso" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;diffplug&#x2F;gitcasso</a><p>Video walkthrough (2 mins of the tool, 12 mins of its development tooling): <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=wm7fVg4DWqk" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=wm7fVg4DWqk</a><p>And a text writeup with timestamps to the video walkthrough <a href="https:&#x2F;&#x2F;nedshed.dev&#x2F;p&#x2F;meet-gitcasso" rel="nofollow">https:&#x2F;&#x2F;nedshed.dev&#x2F;p&#x2F;meet-gitcasso</a>

[Other] Show HN: Static builds of popular open source libraries on npmjs.org

[Other] Show HN: Rustacean AI – Tracking Rust's Expanding Role in AI Rustacean AI, a weekly newsletter exploring how Rust is shaping the future of AI and Machine Learning. Issue #1: From Models to Data — Rust’s Expanding Role in AI Covers the growing Rust ecosystem around AI frameworks (Burn), data systems (Polars, Qdrant, Daft), and emerging tools that power safe, high-performance pipelines. The goal is to curate news, releases, and experiments from across the Rust + AI community — helping developers follow how the language is influencing the next generation of infrastructure. Read first issue here → <a href="https:&#x2F;&#x2F;rustacean.ai&#x2F;p&#x2F;issue-1-from-models-to-data-rust-s-expanding-role-in-ai" rel="nofollow">https:&#x2F;&#x2F;rustacean.ai&#x2F;p&#x2F;issue-1-from-models-to-data-rust-s-ex...</a>

[Other] Show HN: I extracted BASIC listings for Tim Hartnell's 1986 book Tim Hartnell was one of the most prolific authors during the early days of the home computing boom, writing many popular books covering genres of games on different platforms and, in this case, artificial intelligence.<p>I&#x27;ve extracted the BASIC program listings from Hartnell&#x27;s 1986 book &#x27;Exploring Artificial Intelligence on Your IBM PC&#x27; and organized them along with a PC-BASIC runtime environment and instructions so you can try these programs out yourself.<p>Even though the AI landscape has changed enormously since Hartnell first wrote this book, I hope one or two of you will get some value out of these program listings if you&#x27;re interested in exploring the fundamentals of AI on home-computing platforms as they were in the 1980&#x27;s.<p>Tim Hartnell unfortunately passed away in 1991 at the young age of 40, and without his writing I imagine more than a few of us would not have found the start in computing we did. Thanks Tim.

[Other] Datastar: Lightweight hypermedia framework for building interactive web apps

[Other] Show HN: Open source, logical multi-master PostgreSQL replication

[Other] Show HN: Open-Source Voice AI Badge Powered by ESP32+WebRTC hi!<p>video[0]<p>The idea is you could carry around this hardware and ask it any questions about the conference. Who is speaking, what are they speaking about etc... it connects via WebRTC to a LLM and you get a bunch of info.<p>This is a workshop&#x2F;demo project I did for a conference. When I was talking to the organizers I mentioned that I enjoy doing hardware + WebRTC projects. They thought that was cool and so we ran with it.<p>I have been doing these ESP32 + voice ai projects for a bit now. Started with an embedded sdk for livekit[1] that jul 2024 and been noodling with it since then. This code then found its way into pipecat&#x2F;livekit etc...<p>So I hope it inspires you to go build with hardware and webrtc. It&#x27;s a REALLY fun space right now. Lots of different cheap microcontrollers and even more cool projects.<p>[0] <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=gPuNpaL9ig8" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=gPuNpaL9ig8</a><p>[1] <a href="https:&#x2F;&#x2F;github.com&#x2F;Sean-Der&#x2F;embedded-sdk" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;Sean-Der&#x2F;embedded-sdk</a>

[Other] Show HN: Summeze – Turn videos into editable LaTeX summaries in seconds

[Other] Finding a VS Code Memory Leak

[Other] GitHub Issues

[Other] Show HN: A context aware backend for AI coding agents Hey HN, I’m Hang. Today I’m open-sourcing InsForge, a context aware backend for AI coding agents.<p>When using agents like Cursor or Claude to build applications, they often assume what the backend looks like instead of inspecting it. Without access to the actual backend state, they fall back on outdated information, which leads to concrete issues. For example:<p>1. Storage, edge functions, and database logic are closely connected. Without understanding how these parts interact, existing setups get overwritten and important flows break. 2. Database migrations conflict with foreign keys or miss functions because the agent never inspects the live schema. 3. Recreating tables or adding columns that already exist, which leads to conflicts and failed deploys.<p>These problems are not about the agent’s ability to code. They happen because there’s no structured way for the agent to inspect and understand the actual backend before acting.<p>To address this, I built InsForge, which exposes the backend in a structured way and gives the agent direct control:<p>1. Introspection endpoints for schema, relations, functions, triggers, policies, routes, storage, roles, documentation, logs and events 2. Control endpoints for operations usually done through CLI, dashboards or SQL editors<p>InsForge is a full backend platform that includes:<p>- Postgres - Authentication - Storage - Edge functions - Built in AI-model endpoints (via OpenRouter)<p>On top of this, it exposes structured backend metadata and control capabilities through an MCP server and tools, providing a structured, self-describing interface for agents to inspect schemas, policies, triggers, and docs, and interact with the backend.<p>It’s open source and can be self hosted (<a href="https:&#x2F;&#x2F;github.com&#x2F;InsForge&#x2F;InsForge" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;InsForge&#x2F;InsForge</a>) or try our cloud service at (<a href="https:&#x2F;&#x2F;insforge.dev&#x2F;" rel="nofollow">https:&#x2F;&#x2F;insforge.dev&#x2F;</a>).<p>We love feedback!