I'd love for you to check it out, and contribute your own insights and best practices to make this a comprehensive resource for the community.
Cheers!
The idea is simple: instead of setting up runners, servers, or cloud infrastructure, you can just run it on your laptop (or a small server). It integrates with GitHub issues and PRs, automates builds and deploys, and makes it easy to access the running app remotely (e.g. over something like Tailscale).
For me, this solved the problem of quickly testing and deploying code from issues/PRs in a lightweight way. Existing CI/CD systems felt like overkill for that use case.
Repo: https://github.com/simonjcarr/gocd
I’d love feedback from the community — especially on whether this minimal approach to CI/CD is something others would find useful, and what features you’d expect in a tool like this.
— a simple, embedded vector database that stores everything in a single file, just like SQLite.
The problem:
If you’re a developer building AI apps, you usually have two choices for vector search
- Set up a server (e.g. Chroma, Weaviate) - Use a cloud service (e.g. Pinecone)
That works for production, but it’s overkill when you just want to:
- Quickly prototype with embeddings - Run offline without cloud dependencies - Keep your data portable in a single file
The inspiration was *SQLite* during development — simple, local, and reliable.
The solution:
So I built VectorLiteDB
- Single-file, embedded, no server - Stores vectors + metadata, persists to disk - Supports cosine / L2 / dot similarity - Works offline, ~100ms for 10K vectors - Perfect for local RAG, prototyping or personal AI memory
Feedback on both the tool and the approach would be really helpful.
- Is this something that would be useful - Use cases you’d try this for
Everything is purely deterministic, the universe is calculated from SHA-256 hashed seeds using the golden ratio as primordial constant. There’s no database, no pre-saved data, just pure math. Time itself is treated as a coordinate, so the universe exists as a 4D structure where any moment can be computed on demand. Shut it down for weeks, restart, and planets have still been orbiting. Open the same world on multiple devices and you’ll see identical cloud formations, lava flows, even particle effects—always perfectly synchronized (if your clocks are synced).
The simulation applies real physics, Kepler’s laws, tidal locking, Roche limits, hydrostatic relaxation for moons, and orbital temperature variations. Scale is mind-boggling, 300 tredecillion potential planets, far beyond anything that could ever be explored. The backend runs on Python/Flask with Hypercorn, the frontend on React + Three.js, connected via a custom MIT-licensed “vite-fusion” plugin we made. Everything is generated in real time, no storage needed.
The Atlas includes 26+ planet types, fictional elements, moons evolving over geological timescales, and rare life forms that display Arecibo-style messages when analyzed. There’s resource mining and spaceship progression as gamification features. At its core, it’s a playable implementation of Einstein’s block universe theory, all moments exist simultaneously in the mathematical structure, you’re just moving through different temporal slices.
You can try the live demo or run your own universe locally. When installed, you can choose between Core Continuum (a shared seed universe evolving since 1986, my birth year) or Design the Multiverse (your own unique cosmos with a fresh seed). I’d love feedback on the procedural generation algorithms and ideas for expanding the physics simulation!
- GitHub: https://github.com/SurceBeats/Atlas - Docker: bansheetech/atlas:latest - Demo: https://the-atlas.koyeb.app - Alt Demo: https://surcebeats.node1.homedock.cloud:42424
Thanks for reading this far! <3
The problem is that I actually like changing my theme every now and then. And I use a different theme for my terminal app and e.g. VSCode. A single prompt config couldn't fit into everything.
So I built the same prompt from scratch with Swift, and made it automatically adapt to the terminal colors, with proper Lab color interpolation. I used that until a couple of weeks ago, when I wanted pick up Rust. I was quite happy with the hacked-together prompt I had been using for a year, and thought that rewriting it might be a suitable project.
I decided to turn this Rust version into an open source project, as I might not be the only one who'd enjoy a fancy powerline prompt without a screaming color palette.
The source is available at https://github.com/juuso/gigawatt
Last year, I noticed the trend of Digital Marketing on TikTok, and I wanted to give it a try by selling eBooks.
My goal was to sell eBooks that actually help people solve their problems. I opened Canva, started a new design… but quickly realized I had no idea how to write an eBook — and no inspiration at all.
After watching tons of videos on how to make an eBook, I finally finished one… in three months. I started looking for ways to create eBooks faster, but every service I found was way too expensive for what I needed.
So I built QuickTome AI. Now I save hours every week, without breaking the bank.
(1) https://prompt-craft.github.io/ai-study/#sentient-response-p...
I built a simple "AI Story Generator app" to help users create stories, books, novels, or short creative writing pieces with the help of AI.
Features: - AI-powered story generation - Choose genres like fantasy, horror, romance, or sci-fi - Generate stories based on custom prompts - Export and share easily
Why I built it: I noticed many people (especially casual writers and students) struggle with writer’s block or just want to play with creative storytelling. This app tries to make that easier and more fun.
I’d love feedback from the HN community: - Which features would make an AI story app more valuable for you? - Any suggestions for improving the user experience? - Would you prefer more control (plot structure, characters) or more