2025年12月14日 的 Show HN

I created this because sometimes I want more than rlwrap but less than emacs. icl aims to hit that middle sweet spot.

It's a terminal application with context-aware auto-complete, an interactive object inspector, auto-indentation, syntax colouring, persistent history, and much more. It uses sly to communicate with the child lisp process and aims to be compatible with any sly-supporting implementation. I hope others find it useful!

Stepped is a Rails engine, extracted out of Envirobly where it powers tasks like application deployment, that involve complex, out-of-the-band tasks like DNS provisioning, retries, waiting for instances to boot, running health checks and all the fun stuff of a highly distributed networked system.

I built this because I wanted a digital I Ching that honors the original math, rather than just Math.random().

Technical detail for the curious:

The simulation follows the traditional Yarrow Stalk algorithm (Da Yan). The core of this algorithm relies on modulo-4 arithmetic on the stalks remaining after a random split.

To make it physically realistic, I used a Box-Muller transform to simulate the user splitting the stalks with a Gaussian distribution (since humans naturally tend to split near the middle), rather than a uniform random split.

I was worried this 'human bias' might skew the modulo probabilities, so I ran a 1-million-run Monte Carlo simulation to verify.

The Result: The remainders [0,1,2,3] still emerged with a near-perfect 25% distribution (deviation < 0.05%).

So, the app simulates the physics of human imperfection while preserving the mathematical perfection of the probability distribution (where Old Yin is 1/16).

Stack: Next.js + Tailwind css

Happy to answer any questions about the math or the hexagrams!

Tambourine is an open source, fully customizable voice dictation system that lets you control STT/ASR, LLM formatting, and prompts for inserting clean text into any app.

I have been building this on the side for a few weeks. What motivated it was wanting a customizable version of Wispr Flow where I could fully control the models, formatting, and behavior of the system, rather than relying on a black box.

Tambourine is built directly on top of Pipecat and relies on its modular voice agent framework. The back end is a local Python server that uses Pipecat to stitch together STT and LLM models into a single pipeline. This modularity is what makes it easy to swap providers, experiment with different setups, and maintain fine-grained control over the voice AI.

I shared an early version with friends and recently presented it at my local Claude Code meetup. The response was overwhelmingly positive, and I was encouraged to share it more widely.

The desktop app is built with Tauri. The front end is written in TypeScript, while the Tauri layer uses Rust to handle low level system integration. This enables the registration of global hotkeys, management of audio devices, and reliable text input at the cursor on both Windows and macOS.

At a high level, Tambourine gives you a universal voice interface across your OS. You press a global hotkey, speak, and formatted text is typed directly at your cursor. It works across emails, documents, chat apps, code editors, and terminals.

Under the hood, audio is streamed from the TypeScript front end to the Python server via WebRTC. The server runs real-time transcription with a configurable STT provider, then passes the transcript through an LLM that removes filler words, adds punctuation, and applies custom formatting rules and a personal dictionary. STT and LLM providers, as well as prompts, can be switched without restarting the app.

The project is still under active development. I am working through edge cases and refining the UX, and there will likely be breaking changes, but most core functionality already works well and has become part of my daily workflow.

I would really appreciate feedback, especially from anyone interested in the future of voice as an interface.

We're Carolina Cloud - managed data science infrastructure at ~1/3 the cost of AWS.

I left my job earlier this year after watching companies get crushed by cloud bills for workloads that didn't need hyperscaler complexity. Some examples from my previous life: - $1k/month for a basic 16 vCPU VM - $50k/month for a high-RAM instance - Over $1k/month for notebook platform start-stop execution

We built Carolina Cloud for data scientists and small teams who need serious compute without the sticker shock. Our sweet spot: if you're running VMs, notebooks, or RStudio and not deeply tied to AWS/Azure/GCP service ecosystems, we can save you a lot of money.

What we offer: - Standard Ubuntu VMs - One-click Marimo notebooks - One-click RStudio Server and Shiny hosting - S3-compatible object storage (launching soon) - Prepay discounts for commitments as short as 2 weeks - SOC2-certified, HIPAA-compliant datacenter in Charlotte, NC

Simple pricing: $0.005/vCPU/hr, $0.005/GiB RAM/hr, and $0.0001/GiB of hot storage/hr on AMD EPYC Turin processors. A 32 vCPU, 128GB RAM instance runs ~$240/month vs $800+ on AWS.

We're not trying to replicate every AWS service - if you need Lambda + Secrets Manager + S3 with pre-signed URLs, stick with AWS. But if you're a hedge fund running backtests, a biotech team analyzing genomics data, or a researcher who just needs a beefy VM without surprise egress fees, we're 1/3 the price.

Check us out at console.carolinacloud.io - happy to answer questions about our infrastructure, pricing, or why we think there's room for regional clouds built on owned hardware.

I've worked in tech my entire career, and over time mindfulness meditation has become less of a "wellness habit" and more of a practical tool for keeping my mind clear.

I encourage mindfulness meditation to everyone because it's that impactful on stress and awareness in everyday life.

After years of practice, I've that most meditation apps eventually became distracting or costly.

To me, mindfulness apps that provide guidance, content, forced streaks, and incentives seem to replace the practice of mindfulness itself with some gamification of the act.

So I built Center, a silent meditation timer with a few constraints:

• No guided sessions, voices, or music • No subscriptions, ads, or paywall • Optional tracking, no gamification • Works across iPhone, Apple Watch, iPad, and Mac via iCloud • Donation-supported, with no features locked

The goal has never been growth or engagement, but reliability. I wanted something closer to a clock than a coach.

This is a side project, but it’s been interesting to see how a deliberately "boring" tool changes how I personally practice mindfulness, and how that carries over into my work (especially decision-making and stress tolerance).

Site: https://centertimer.com

I've been building side projects for over a decade. Most failed to find Product‑Market‑Fit - I know the frustration of shipping features that don't stick.

The painful truth I learned: many teams chase features over feedback. Worse, most feedback is shallow or useless because people avoid being critical.

So I built user.mom to fix the process, not just add another tool. It maps the full PMF journey: Landing Pages to validate demand, Surveys to gather structured signals, Feedback Boards to organize requests, Customer Voting to prioritize, and Integrations (CSV, webhooks, API) to scale what works.

If you’re tired of guessing what customers want, start your PMF workflow - the first product is free for every user.

I built Read Fast after realising I was saving hundreds of things to read and reading almost none of them.

It automatically turns saved content into a scheduled briefing and sends it to my inbox at whatever time I choose.

New version since last post actually allows you to bulk upload URLs. So you can just dump links here and Read Fast will scrape and brief for you.

Great for actually reading the things I want to. Curious if that resonates with people?

I love the look of HN on Christmas Day, but I never feel like I get quite enough of it in just one day. To rectify this catastrophic problem, I created a browser extension to give me the Christmas HN experience from the day after Thanksgiving until the first work-day of the new year.

I also added a fun "extra festive" mode that enhances the festivity level of the site beyond it's normal look. I love it, but I expect it to be controversial :-D

It's pretty easy to `git clone` and install it (instructions for Firefox and Chrome in the README.md). I haven't put it in the extension stores as I have no idea if anybody else would want it, but if it turns out there are dozens of us[1] that want to use it, I will do so.

Goal is to fully support Firefox and Chrome (as I'm a Firefox user primarily), though I haven't found a way to get the Christmas tree animated gif to load in Firefox so that one part doesn't currently work. I tried all kinds of hacks to convert the tree into an animated svg, and even generate new ones, and they are laughably horrible. I'll continue to look and try stuff but for now at least it will just display the alt text on Firefox.

Settings are accessible by clicking on the extension in the menu or tray.

[1]: A nod to the iconic Dr. Tobias Fünke from Arrested Development

PhotoToVideoAI is an AI-powered tool that turns your photos into dynamic videos. Upload a photo and a prompt, and in about 30 seconds you'll get a high-quality video in resolutions up to 1080p and durations of 5 or 10 seconds. It’s designed for content creators, marketers, and photographers, Feedback is welcome!

Powerlifting programming isn’t as complicated as spreadsheets make it look; once you strip away the noise, the structure fits comfortably on a phone. I think of it as two parts: plain sets/weight/reps/RPE/1RM, and the custom analytics you build on top. SigmaLifting handles the first part entirely on your phone with a mobile-first UI and a flexible data model that can express most spreadsheet-style programs. For the second part, you export the data and run it through your own tools (or an LLM) for whatever analysis you want.

The biggest problem with Tic-Tac-Toe is that it almost always ends in a draw. Tic Tac Flip tries to fix that!

Learn the rules in Learning Mode or below:

- Winning Criteria: 3 Ghosts (Flipped O or X, which can be a mixture). It's not just 3 Os or 3 Xs anymore!

- Flipping Mechanic: When one or more lines having only O and X are formed, the minority of either all Os or all Xs get flipped to a Ghost, and the majority gets removed from the board. E.g., A line of 2 Os and 1 X leads to 1 X ghost and the removal of 2 Os.

- Active Flip: You can actively flip your O/X to a Ghost (or flip a ghost back) once per game.

- Placing Ghost Directly: You can place a "Ghost" piece directly as a final winning move (only once, and only when there are two existing ghosts in a line).

I'm looking for feedback on the game balance and learning curve. Specifically: - Is the "Ghost" and "Flip" mechanic intuitive? - Is the Learning Mode helpful? - Is the game fair? Any rule adjustments needed? - Any bugs or issues?

Any suggestions or comments would be much appreciated. Thank you in advance!

I built depup, a CLI tool to scan Python dependencies, check PyPI versions, classify upgrade impact, and support CI workflows.

Docs: https://saran-damm.github.io/depup/ Repo: https://github.com/saran-damm/depup/

g023's OllamaMan - Ollama Manager OS style GUI management of ollama Server - Open Source LLM Management using PHP/JS/Sqlite and a web browser. Integrated apps for chat, terminal viewer, model management (can pull or delete; huggingface gguf supported too). Advanced model creation now implemented. Supports images in chat and speech to text/history/etc. Lots of quick access buttons, and auto sets itself up if db doesn't exist. Adding more bells and whistles as I go. Not meant for a public facing folder, so protect as you see necessary. Open source, BSD 3-Clause so have fun and thanks for giving it a glance. I just made it yesterday, so I'm sorry if it has some rough bits as I'll fudge those into place hopefully shortly (menus on the topbar/settings tabs). Right now fully functional.

built a CLI tool that scans your project dependencies and automatically creates specialized documentation agents for Claude.

How it works: 1. Run npx claude-docs in your project root. It generates a /sync-docs command for Claude Code. 2. Run claude /sync-docs 3. You get a personalized /research slash command configured specifically for your project's stack and @ specialized subagents for the different dependencies

This is my first open-source project.

Looking for feedback.

Repo: https://github.com/ii-vo/claude-docs

Demo: https://www.claudedocs.com/

Hello HN,

I built "Wax On, Python", a browser-based Python 'dojo' that focuses on syntax precision through repetition and muscle memory.

https://waxonpython.com

The Problem: When learning to code, it's easy to copy-paste or gloss over syntax. Many tutorials explain the logic but don't force you to actually type the code. Sometimes beginners encounter basic syntax errors (missing colons, wrong indentation, mismatched braces etc) once they graduate to script writing.

The Solution: Inspired by a classic fictional tale, this app takes an approach as if lessons are given by a 'strict sensei': - Drills: You are given code snippets and must type them exactly character-for-character. - Immediate Feedback: Typo? 'FOCUS!' 'SLOPPY.' Correct? 'GOOD.' - Muscle Memory: starting with simple print statements and building up to loops, classes, and eventually pandas dataframes. - Philosophy: Between drills, you get philosophical Sensei wisdom explaining the why behind the concepts (eg: 'Variables are buckets', 'Indentations are choices' etc).

The tech... - Runs entirely in browser. Code executes locally so no servers or latency. - Runs a Pyodide Python REPL in-browser so user can see real errors as they make them. Includes Pandas package for data science drills. - Stack: vanilla js, HTML, css.

Why I made it I run a Python Learning Group at work and wanted a good engaging starting point for anyone who is curious about Python regardless of their professional or technical background - the goal being to make syntax automatic so you don't have to think about it, freeing the mind for higher-level logic, and giving people a sense of the learning journey they are about to embark on.

Any feedback on curriculum pacing and the "tough love" approach would be helpful - does it work in your case? would love to hear it.

Note: the app was actually blocked at work due to strict wasm rules implemented in Zscaler :(

Hey HN!

I kept noticing great posts dying in /new while some mediocre ones hit front page at the right time. So I built a tool that analyzes your post and suggests optimal posting windows based on historical HN data. Also if provides general feedback based on data + similar posts.

From the tech side of it, both the fe and the HN API are open source here:

- Frontend: https://github.com/desplega-ai/ai-toolbox/tree/main/willitfr... - SQL API: https://github.com/desplega-ai/ai-toolbox/tree/main/hn-sql

The backend is done with a simple python script that fetches data into parquet files, and then queries them using duckdb, which work fine so far. The latencies go from ms to few seconds depending on the query.

PS: I haven't finished the full sync yet (~85% done).

Create stunning AI videos with synchronized audio using Seedance2 AI Video. The latest AI video generation tool that adds sound effects, dialogue, and ambient noise.

Basically, paste any job URL and it automatically researches the job posting to draft personalized cover letters and resume tips.

You get to watch the AI research the role like a human would in real-time.

trylockedin.app

Hi HN

I built a Hazbin Hotel OC Maker, an AI-powered tool that helps fans create original characters (OCs) inspired by the Hazbin Hotel universe.

As a fan myself, I noticed that many creators struggle to turn vague ideas into a complete OC — especially when it comes to personality, backstory, visual traits, and tone consistency with the Hazbin Hotel style. Most tools focus only on images or only on text, so I wanted to combine both into a single workflow.

What it does

Generates original Hazbin Hotel–style characters (not existing canon characters)

Helps define:

Name & demon type

Personality traits

Backstory & motivations

Visual description suitable for artists or AI image tools

Designed for writers, roleplayers, and artists

Works entirely in the browser, no setup required

Why I made it

I’m building a broader AI OC creation platform, and this is one of the first focused generators aimed at a specific fandom. Hazbin Hotel has a very distinct tone and aesthetic, which made it an interesting challenge for prompt design and style control.

Link

Hazbin Hotel OC Maker https://aiocmaker.com/oc-maker/hazbin-hotel-oc-maker

Notes

This is a fan-made tool and is not affiliated with Vivienne Medrano or A24

Feedback is very welcome — especially from writers and artists who create OCs regularly

Happy to answer any questions about the tech, prompts, or design decisions. Thanks for checking it out!

— An indie builder