매일의 Show HN

Upvote0

2025년 7월 25일의 Show HN

19 개
337

Price Per Token – LLM API Pricing Data #

pricepertoken.com faviconpricepertoken.com
129 댓글12:39 PMHN에서 보기
The LLM providers are constantly adding new models and updating their API prices. Anyone building AI applications knows that these prices are very important to their bottom line. The only place I am aware of is going to these provider's individual website pages to check the price per token.

To solve this inconvenience I spent a few hours making pricepertoken.com which has the latest model's up-to-date prices all in one place.

Thinking about adding image models too especially since you have multiple options (fal, replicate) to use the same model and the prices are not always the same.

113

The Montana MiniComputer #

mtmc.cs.montana.edu faviconmtmc.cs.montana.edu
22 댓글2:25 PMHN에서 보기
Hey HN, we just released the 1.0 of the MonTana Mini Computer (MTMC-16), a virtual teaching computer to help students understand how low level computing works. It is a 16 bit computer with only 4k of ram, but we've made some design choices that help maximize what you can accomplish with the limited hardware.

It is written in java (sorry) and provides a web interface that has:

- a blinken-lighten display for registers

- a memory view with different filters you can apply

- a Gameboy-like game pad

- a console you can use to interact with the computer (including running assembly instructions directly)

- a file browser with an integrated editor for editing file

So everything you need to get going on low level programming.

It includes some sample code, including snake and conway's game of life, in the /src directory.

You can watch a quick start video here: https://www.youtube.com/watch?v=m_6pZ_sT3y0

We have the start of a C compiler for the machine, but that's still a work in progress. We plan on improving the interactivity and visual feedback over the next few months, so any feedback you can give us would be very much appreciated!

56

A macOS clock that stays visible when coding or binging in fullscreen #

cornertime.app faviconcornertime.app
42 댓글4:02 PMHN에서 보기
Have you ever wished you could see the time while coding or watching videos in fullscreen or when your menu bar is hidden? Corner Time is a minimalist macOS app designed for those who value a clean desktop but still want the clock to be always visible.

Unlike other widgets or solutions, Corner Time sits unobtrusively in the corner of your screen as if it were a native part of macOS—even when you’re in fullscreen mode or using auto-hide for your menu bar. The app is highly customizable: you can choose your preferred date format, adjust the style, match your system language, and use it seamlessly across all your monitors.

Corner Time has quickly become a favorite among digital minimalists and power users. Many users describe it as a long-awaited solution, praising its thoughtful design.

Feedback and suggestions are welcome!

39

I built a biological network visualization tool #

nodes.bio faviconnodes.bio
28 댓글11:38 AMHN에서 보기
I've been working on nodes.bio - an interactive tool for visualizing biological networks and systems thinking. The tool features interactive network visualization powered by Cytoscape.js, with real-time graph editing and manipulation capabilities. It supports JSON import/export and provides a responsive design that works seamlessly on the desktop (mobile-friendly version coming later).

The tech stack combines modern frontend technologies with robust backend architecture. The frontend uses Next.js 14 with TypeScript and Cytoscape.js for the visualization engine. The backend is built with FastAPI and Python.

The featured demo showcases a Traumatic Brain Injury Nasal Spray mechanism of action visualization, demonstrating the tool's capability to handle complex biological pathway mapping.

You can explore the live demo at <https://nodes.bio> to see the TBI Nasal Spray visualization in action, along with other biological network examples.

I'd love feedback on the visualization capabilities or any suggestions for biological data integration. What do you think?

5

BlackMagic-JS – Automatic dark mode framework that just works #

github.com favicongithub.com
1 댓글6:29 PMHN에서 보기
I've been frustrated with dark mode implementations for years. Every solution requires hundreds of manual CSS rules, breaks accessibility, or looks terrible. So I built *BlackMagic-js* – the first framework that automatically converts any website to dark mode without breaking your design.

## The Problem

Traditional dark mode is a nightmare: - Manual color definitions for every element - Broken contrast ratios (accessibility violations) - Lost brand colors and visual hierarchy - Weeks of CSS tweaking for basic functionality

## The Solution

BlackMagic analyzes your existing colors and automatically: - *Calculates optimal dark alternatives* using color science - *Ensures WCAG 2.1 compliance* (4.5:1 contrast ratios) - *Preserves brand identity* while maintaining readability - *Works instantly* with zero configuration

## How It Works

```javascript // Traditional approach: 500+ lines of CSS .dark-mode .button { color: #fff; background: #333; } .dark-mode .text { color: #e0e0e0; } // ... endless manual definitions

// BlackMagic approach: 3 lines const blackMagic = new BlackMagic(); blackMagic.toggle(); // Everything just works ```

*Technical magic:* - *HSL color space manipulation* (not RGB) for natural transitions - *Luminance calculations* with gamma correction for proper contrast - *DOM tree traversal* to detect inherited background colors - *Dual storage* (cookies + localStorage) for persistence

## Real Performance

Tested across 50+ websites: - *98.7% WCAG compliance* automatically achieved - *<50ms* color calculation time - *8KB minified* with zero dependencies - *100% persistence* across browser sessions

## Quick Start

```bash npm install blackmagic-js ```

```javascript import BlackMagic from 'blackmagic-js';

const darkMode = new BlackMagic({ backgroundColor: '#1a1a1a', factor: 0.4 // Adjustment intensity });

// Add to any button button.addEventListener('click', () => darkMode.toggle()); ```

CDN version: ```html <script src="https://cdn.jsdelivr.net/npm/blackmagic-js@latest"></script> ```

## Why This Matters

Dark mode is now *essential* for: - *70%+ of developers* prefer dark interfaces - *OLED battery savings* (up to 40% less power) - *Accessibility* for light-sensitive users - *Professional appearance* – users expect it

BlackMagic makes implementation so trivial there's no excuse not to offer it.

## Advanced Features

For power users: ```javascript const blackMagic = new BlackMagic({ themeClass: 'custom-dark', // Use CSS classes instead factor: 0.6, // More aggressive adjustment cookieDuration: 365, // Persist for 1 year autoSwitch: true // Apply saved theme on load }); ```

## Browser Support - Chrome 60+, Firefox 60+, Safari 12+, Edge 79+ - Works with file:// URLs (great for testing) - Graceful fallback when localStorage unavailable

## Open Source & Testing

- *GitHub*: https://github.com/LucAngevare/BlackMagic-js - *NPM*: https://www.npmjs.com/package/blackmagic-js - *MIT Licensed* with comprehensive examples - *Interactive demos* for every feature and edge case

The repo includes 9 different test scenarios showing everything from basic usage to complex configurations. You can literally see it work on any website in seconds.

3

LogMerge – View multiple log files in a merged view #

github.com favicongithub.com
0 댓글10:53 PMHN에서 보기
Hey HN!

I needed a tool to view multiple log files in a merged view, and easily filter based on the specified fields. Spent a good amount of time searching, but couldn’t find any open source tool that quite did what I wanted. So, ended up building a custom solution instead (I would appreciate suggestions on tools that have similar functionality).

I don't know much about GUIs (most all my PC based utilities are CLI) - but I did have the following:

  - I know enough Python to spot obviously wrong things

  - Some knowledge of how to make programs performant in general

  - ... and tokens to burn :)

GitHub : https://github.com/faisal-shah/pylogmerge Usage Video: https://youtu.be/37V_kZO2TLA

Key Features:

  - Merge and display multiple log files in a single, chronologically ordered view

  - Live log monitoring with auto-scroll

  - Add files individually or discover them recursively with 
regex filtering - Plugin-based system to support any log format (easy to extend!)

  - Filtering: discrete values, numeric ranges, regex/text, and time-based queries

  - Color-coded file identification

  - Configurable columns and ordering

  - Built-in plugins for syslog, CANKing (CAN Bus monitoring tool), and another custom log format called dbglog.

If you have any feedback or questions, let me know! Hope someone else finds it useful.
3

Narev – Open-Source FinOps for AI and Cloud Spend #

github.com favicongithub.com
2 댓글3:28 PMHN에서 보기
Hey HN,

I've been working on unifying SaaS billing data using FOCUS 1.2 from the FinOps Foundation. Started with cloud providers (AWS, Azure, GCP) and OpenAI.

Narev ingests billing data, normalizes it with FOCUS 1.2, and gives you one dashboard for everything. Self-hosted so your data stays private.

Try it:

docker run -d \ --name narev \ -p 8000:8000 \ -v $(pwd)/data:/app/data \ -e DEMO="true" \ ghcr.io/narevai/narev:latest

Then hit http://localhost:8000

This is v0.1.0—stage, early but functional. Looking for feedback on which SaaS integrations to prioritize next, bug reports, and contributors.

GitHub: https://github.com/narevai/narev

3

Metaessen 2.10 – Manage resources for practical use #

metaessen.com faviconmetaessen.com
2 댓글3:56 PMHN에서 보기
Hi everyone,

Metaessen is a resource manager, and manages resources for practical use. It is practical, safe, concise, easy to use, and open. It can be used as a better bookmark manager or note-taking app.

For more information, please refer to the website: https://metaessen.com/

I just released the version 2.10.

Highlights or big features on the version:

- Group selection and deletion - Shortcuts for creation, selection, or other actions. - System tree condition to searcher, keyword filter - Options to delete the corresponding resource(s) when deleting instance(s) or a system - Options to delete system tree, descendants or only descendant instances when deleting a system - Improved UI and user experience - Display data statistics about total number of entities

For more, please refer to: https://metaessen.com/docs/blog/release-v2.10

If you have any issue or question, please contact me via the email on the home page.

Thank you

3

I built a CSV/XLSX editor that uses JavaScript to manipulate data (OSS) #

fileglance.info faviconfileglance.info
0 댓글5:01 PMHN에서 보기
Hi everyone,

I work in enterprise IT, handling diverse data exports from various systems/APIs.

Frustrated by:

1. The need for different tools based on file formats.

2. The lack of tools optimized for quickly understanding data.

3. Messy files often need to be cleaned before use.

... I built my own solution as a side project and a fun way to learn React and Tailwind.

Maybe it helps others as well.

It aims to be both:

- Simple: Just drag and drop a file; it automatically detects encoding, delimiter, headers, etc.

- Powerful: Run arbitrary JavaScript to filter and transform data at scale.

Try it out: https://www.fileglance.info/

Source code: https://github.com/dell-mic/file-glance

I’d love to hear your feedback!

2

StackDAG: Build and share application stacks as DAGs #

stackdag.pages.dev faviconstackdag.pages.dev
0 댓글3:00 PMHN에서 보기
I’ve just launched StackDAG into public beta. It’s a platform for building and sharing application stacks using Directed Acyclic Graphs (DAGs).

You can model anything from a classic PERN stack to your own infrastructure layout, using a flexible visual editor. It includes: - Templates for common stacks - A wide component library - Forking and sharing - Import/export as JSON - Community feed for trending DAGs

The idea is to make stack design more modular and shareable, especially for teams and devs working across multiple tools.

I’m very open to feedback at this stage, and there will be frequent updates during beta and beta testers will keep their role (plus early access to premium features later).

Try it here: https://stackdag.pages.dev

Discord: https://discord.gg/VqwqHmg5fn

2

Configurable parser from binary message to JSON (and viceversa) #

gitubo.github.io favicongitubo.github.io
0 댓글3:46 PMHN에서 보기
I developed it because I was tired of writing custom binary parsers for network protocols and wanted to create an easy-to-use tool directly from format experts without the need for developer support - while ensuring adequate performance. Ideal for centralising message encoding/decoding operations in different formats.

GitHub: https://github.com/gitubo/bixit

What binary formats do you deal with regularly? Would something like this be useful?

2

Add viral TikTok audio to work meetings #

soundboard.recall.ai faviconsoundboard.recall.ai
0 댓글3:14 AMHN에서 보기
as title says

i mostly did it because I had a free afternoon at work, built something fun and dumb and it ended up being more dumb and fun than i thought

submit your meeting link and the bot should join the call. users can send commands in-chat and the bot will play the audio

no sign-up required and i’m hosting it for free. hope you have as much fun as i did

cheers!

1

AI image generator with 6 artistic mentors for better prompts #

createvision.ai faviconcreatevision.ai
0 댓글3:16 AMHN에서 보기
Hi HN! I built CreateVision AI, an image generator that uses AI mentors to help write better prompts.

Link: https://createvision.ai

*What makes it different:* - 6 specialized AI mentors (Renaissance, Anime, 3D, Fashion, Abstract, Traditional Eastern) - Each mentor enhances prompts in their artistic style - Dual engines: Flux Dev + GPT-Image-1 - 2-4 second generation time - 8 language support

*Technical details:* - Built with Next.js 14, TypeScript, Supabase - Custom caching layer for speed - Global lock mechanism for request handling - Vercel Blob for optimized image delivery

*Why I built this:* After analyzing 50k+ generations, I found users averaged 3-4 attempts to get good results. The problem wasn't the AI model - it was prompt writing. Our AI mentors solve this by teaching artistic principles while enhancing prompts.

Free tier: 20 images/day with watermarks Premium: $9/month for 100 images + both engines Ultimate: $15/month unlimited

Would love feedback on the mentor system and any suggestions for improvement!