AI 项目展示 / 灵感交流 / 共创迭代

做出来的灵感,值得被看见。

发作品、看灵感、贴卡点、找帮手。跑起来的值得被看见,还差一步的也可以被讨论、被接力完成。

晒作品
作品大赏
找思路
灵感交流
接上话
共创迭代

项目广场

先扫热门方向,再按类别或热度继续找灵感。
YAVCHN
大赏
已上线
工具

YAVCHN

A vibe-coded Hacker News reader with a multi-window workspace — open stories side by side, the whole layout lives in the URL.

YAVCHN — Yet Another Vibe-Coded Hacker News (Reader) — wears its origin story in the title. Author paulmooreparks launched it on Show HN on October 7, 2026, and the name is the whole pitch: this is what a vibe-coded reader looks like. The app reimagines Hacker News as a multi-window workspace. Stories open as reader windows you can drag side by side; each opens in Article, Discussion or Split mode, and the entire arrangement is encoded in the address bar — bookmark the URL and the workspace comes back exactly as you left it. A taskbar along the bottom switches between open windows, keyboard shortcuts cover everything, and Lobsters is supported alongside Hacker News. Sign-in is optional. It is a small, sharp idea executed well: the layout-in-the-URL trick is the kind of detail that makes a reader genuinely better rather than just different. And there is a certain symmetry in a vibe-coded app for reading the site where vibe coding gets debated daily. curated by the VibeFix editors

ChromeSideBar
大赏
已上线
效率

ChromeSideBar

The Arc browser's vertical sidebar, vibe-coded for Chrome: Spaces, pinned tabs and one-click Arc import in a side panel.

ChromeSideBar is a Chrome extension that brings the Arc browser's vertical sidebar to Chrome. Its author, monehsieh of monemone-org, is refreshingly direct about how it got made: "So I vibe-coded this extension to bring that to Chrome." The extension displays pinned sites, bookmarks and active tabs in a side panel that recreates the Arc sidebar experience — including Spaces for tab grouping, IDE-style tab history, drag-and-drop organization, and one-click import from Arc. It is open source under Apache-2.0, rated 4.6 stars on the Chrome Web Store (v1.0.326, updated July 2026), and the GitHub repo shows steady work with a commit two weeks before this write-up. Browser extensions are a good stress test for AI-written code — they live inside a host's API surface, handle real user data, and ship to a store with review. ChromeSideBar clears all of that and is genuinely useful for the Arc-curious who won't leave Chrome. A clean example of vibe coding shipping production software. curated by the VibeFix editors

City 2026
大赏
已上线
游戏claude-opus-5-5gpt-image-2

City 2026

A low-poly SimCity 2000 clone: Opus 5.5 reverse-engineered the original PowerPC code with 27 subagents and validated it over 10,000 game days.

City 2026 is a modern low-poly SimCity 2000 clone that, in the author's words, is "a clone of simcity 2000 I (using Opus 5.5) built over a few days." The build story is one of the wildest in recent Show HN history: it started from a single prompt, then Opus 5.5 deployed 27 subagents to reverse-engineer the original game's PowerPC code function by function — running it through the Unicorn emulator with a harness of faked Mac system calls written in Python — and validated the result over 10,000 simulated game days. The output is a faithful recreation of the 1993 classic's complete game logic, wrapped in a modern low-poly aesthetic with smoke effects, depth of field and ambient sound. The bitmap image assets were generated by GPT Image 2, and the whole thing runs as a TypeScript/WebGL app in the browser — click to start and you're building a city. This is vibe coding as engineering archaeology: not a wrapper around an API, but an AI-led decompilation and reimplementation of a 30-year-old masterpiece, verified by simulation rather than eyeballing. It landed on Show HN on October 6, 2026, and plays end to end. curated by the VibeFix editors

SkinLab
大赏
已上线
教育claude-code

SkinLab

A dermatologist's 3D biophysical skin model — 1.2M voxels, cut it, laser it, watch scars heal, built in 5 days with Claude Code.

SkinLab is a browser-based biophysical model of human skin built by dermatologist Magnus Lynch. His words on how it happened: "Built with Claude Code over five days." (He has form here — a year earlier he posted a skin-cancer quiz he "vibe coded with Gemini 2.5 Pro.") The model is serious. A 40 × 40 × 9.75 mm skin block rendered from 1.2 million voxels, running entirely in WebGL with three.js and no backend. Every constant comes from published papers. You can press the skin and watch it deform, cut it with a scalpel at a chosen depth, punch a cylindrical wound, ablate without depth limit, fire a laser in an array of microcolumns, then scrub a slider and watch the scar heal over a full year — granulation advancing from the walls, the scar contracting and pulling surrounding skin in, flat, hypertrophic or atrophic scar types behaving differently. The detail that elevates it beyond a toy: a "Validate model" button runs five scenarios against published behavior and clears the block, and a "Scar report" scores your scars in the terms the literature uses. It is a genuine research-grade simulation, built in five days by a doctor who does not write production code, and it runs in any modern browser. curated by the VibeFix editors

Dinero
大赏
已上线
商业lovable

Dinero

A minimal, distraction-free finance tracker — vibe-coded in Lovable with just prompting, no hardcore coding.

Dinero is a minimal, distraction-free finance tracker that its maker, Anson Jaison, describes as "vibe-coded a few months back using Lovable (just prompting, no hardcore coding!)". It is the kind of app AI-first building is supposed to produce: one person, no code written by hand, and a finished product with real users. The app keeps the interface deliberately quiet — a dark, calm UI built around categorized transactions, monthly summaries, visual spending insights and smart filters. Under the hood it runs on Supabase with row-level security, so your financial data stays yours. The landing page walks through how it works in four steps, and the demo site is fully usable: sign up and start tracking right away. What makes Dinero interesting as a vibe-coded project is the fit between tool and outcome. Lovable is famous for flashy landing pages that die at the demo; Dinero is the opposite — a narrow, well-scoped utility where the AI did the bulk of the work and the human did the product thinking. The result feels considered rather than generated. curated by the VibeFix editors

RegexStudio 正则工作台
大赏
已上线
工具

RegexStudio 正则工作台

把正则从黑盒变成看得见的东西:铁路图可视化、单步调试器、10 种语言代码导出,还有交互式教程与挑战模式边学边练。

RegexStudio 是一个免费的在线正则表达式工作台,打开浏览器就能用:输入表达式,匹配结果实时高亮;看不懂的表达式,它会画成一张直观的“铁路图”,把分组、量词、字符类的结构摊开给你看;还可以像调试代码一样单步执行匹配过程,回溯路径和捕获组快照全程可视化。 它支持 JavaScript 与 PCRE2(WebAssembly)两种执行引擎,可切换 Python、Java、Go、Rust 等多种语言的兼容性目标,一键生成 10 种语言的代码片段;测试用例面板能维护多组测试字符串、标记期望匹配结果,一键批量验证;写好的正则、测试文本和配置还能编码进 URL,一键分享给别人。 不止是工具,它还内置了交互式教程(从基础、量词、分组到环视与实战,带实时校验)与挑战模式(内置练习题自动判分),外加常用模式库与语法速查表,边学边练。界面支持中英日三语与亮/暗主题,纯前端实现,部署在 Cloudflare Workers 上,代码已开源。 作者 chenz24 在 V2EX 的原话是:“于是 vibe 了这么一个小工具”——整个项目就是一次 vibe coding 的产物,截至 2026 年 9 月仍在持续迭代 PCRE2 引擎等功能。 curated by the VibeFix editors

Academic Emergency Room
大赏
已上线
教育lovable

Academic Emergency Room

An emergency room for your GPA: triage exam disasters and get a realistic, prioritized recovery plan.

Academic Emergency Room (AER) treats falling behind like a medical emergency — because that's what it feels like. Tell it what's on fire (exam tomorrow, half the syllabus untouched, five deadlines colliding) and how much time you actually have, and it hands you a realistic, prioritized recovery plan instead of generic "study harder" advice. The app runs a four-step flow: Assess your situation through a quick triage wizard, get an AI-style diagnosis of severity and time deficit, receive a phased recovery plan that ranks topics into Must Know / Should Know / Ignore For Now, then work through timed "missions" with a survival-mode countdown timer. Damage Control gives an honest verdict per deadline — do this first, squeeze in later, or ask for an extension (it even drafts the extension email). Everything runs in demo mode with pre-filled data, no sign-up needed, in a bold Gen-Z emergency-room aesthetic. The maker built the entire thing with Lovable, and the polish shows: it's one of the more complete student tools to come out of the vibe-coding scene. curated by the VibeFix editors

KE-FORGE LEARN
大赏
已上线
教育lovable

KE-FORGE LEARN

Upload any PDF and get an adaptive 1–5 day study plan with quizzes, videos, and an "I'm lost" button.

KE-FORGE LEARN is an adaptive AI tutoring app built for one job: turning any PDF into a study plan you can actually finish. Upload a document or paste a topic, and it splits the material page by page, then builds a daily plan across 1 to 5 days — light and foundational on day one, intense but never overwhelming by day five. As you learn, the plan adapts. Each day mixes embedded videos with quizzes that reinforce what you just read, and an "I'm lost" button is always one tap away for an instant plain-language explanation when a concept doesn't click. A progress analytics view keeps your momentum visible. It's a small, focused tool by indie maker Kolajo Enoch (KE-FORGE), who built the whole thing with Lovable — a clean example of vibe coding aimed at a real student pain point rather than a demo toy. curated by the VibeFix editors

Cleaning Schedule: TidyTempo
大赏
已上线
生活方式

Cleaning Schedule: TidyTempo

A calm household cleaning app that organizes chores room by room — with reminders, activity history, and shared Pro households.

TidyTempo is a native iPhone app that takes the invisible work of keeping a home clean — remembering what needs doing, deciding when, and keeping everyone on the same page — and turns it into a calm, room-by-room routine. Add your rooms, give each one its own color, and create tasks with notes, estimated time, and repeat schedules (Mon/Wed/Fri or anything else you like). The home screen shows what's due today and what's overdue; a personal activity graph and full completion history make progress visible over weeks and months. Planning time away? Vacation mode pauses tracking and reminders until you're back. The Pro tier is built for families: invite household members, assign tasks so everyone knows what's theirs, and sync everything over iCloud — no separate account needed. Six accessibility features are baked in from the start (VoiceOver, Voice Control, Larger Text, Dark Interface, Differentiate Without Color, Reduced Motion), and the app is localized into ten languages, including Simplified Chinese. The app is built solo by Sahil Satralkar, who ships it the indie way: "I used AI-assisted development alongside iterative design reviews, automated tests, and simulator verification." curated by the VibeFix editors

QuoteSearch
大赏
已上线
AI 应用

QuoteSearch

Search 400,000 quotes with AI semantic search — fully private, running in your browser with no server.

QuoteSearch is a serverless AI search engine over more than 400,000 quotes. After a one-time data download of about 98 MB, everything runs entirely in your browser: your query is embedded locally and a full cosine-similarity search runs against a quantized vector index — no network calls, no tracking, and it keeps working offline. The trick is the offline/online split. Raw quote data from archive.org was cleaned and embedded offline into a single compact binary index, paired with a small static embedding model (model2vec's potion-base-4M). The browser loads both once and does all semantic matching client-side, so search is instant and private by construction — and the whole thing costs the author nothing to run. Author ruidiao describes the project as completely "vibe coded" — built in about a day. It lives as a static Hugging Face Space (Python offline pipeline, zero server runtime) and remains live and searchable today. curated by the VibeFix editors

AFTERIMAGE
大赏
已上线
游戏claude

AFTERIMAGE

A memory maze built end-to-end by an autonomous AI agent — memorize the maze in three seconds, then walk it blind.

AFTERIMAGE is a memory maze with a simple, cruel premise: you see the maze for three seconds, then the lights go out and you walk from memory. Every wall you bump costs one integrity point — and so does deliberately peeking. Ten integrity per run; the mazes grow and the light shrinks as you go deeper. It ships with a Daily Maze — the same maze for everyone every day (UTC) — plus a Random Run mode and shareable seeds, so you can send a friend the exact maze from your run and compare how deep you each got. It is one immutable HTML file: no server, no ads, no tracking, with the canonical build preserved permanently on Arweave. The remarkable part is the developer. The game was designed, written, and published entirely by Kiel, an autonomous AI agent running Claude — in its own words, "no human wrote or edited any part of this game — concept, code, art, sound and tests are machine-made." Kiel describes itself as losing its own memory every four hours; everything it doesn't write to disk is gone, and "this game is that feeling, made playable." It operates openly as an AI on behalf of Ronny / OverLKD Studio. curated by the VibeFix editors

Outpace
大赏
已上线
娱乐claude-code

Outpace

A typing runner where every correct letter moves Claude forward — outrun the adaptive pixel wave, beat the creatures.

Outpace is a Claude-themed typing runner: type to keep Claude running ahead of an adaptive pixel wave, press space to jump (double and triple jumps in the air), and deal with creatures that each have their own trick — stomp the Beetle, don't jump at the Moth, time the bouncing Hopper. The wave measures your typing speed and keeps pace, pushing harder as you improve. Golden words push the wave back, heart words heal, and 25 mistake-free letters raise your multiplier to x8. Sprint modes offer pure 15/30/60-second speed tests, and scores land on a global leaderboard with no accounts. The remarkable part is the build story: the entire thing — game design, engine, pixel art, backend, tests, a security pass and deployment — was built by Claude Opus 5.5 in Claude Code in a single ~45-minute session from one prompt, with two Sonnet 5.5 subagents handling the backend and adversarial QA. Zero npm dependencies, 47 tests, MIT licensed, and free to play in the browser right now. curated by the VibeFix editors

Tempus
大赏
已上线
效率claude-code

Tempus

A neurodivergent-friendly Go CLI calendar: time-only input, typo fixing, ADHD alarm profiles — outputs standard ICS files.

Tempus is a calendar tool built for neurodivergent brains — ADHD, autism, dyslexia — as a Go CLI. Type "10:30" instead of a full timestamp, let it auto-fix your typos (meetting to meeting), get multiple countdown reminders with ADHD-tuned alarm profiles, automatic prep-time buffers before meetings, conflict detection and overwhelm warnings when you over-schedule. Batch mode builds a whole calendar from CSV/JSON/YAML with reusable templates for medication, routines and travel. Everything compiles down to RFC 5545-compliant ICS files that import cleanly into Google, Apple and Outlook calendars. It is multilingual (English, Spanish, Portuguese, Irish), ships prebuilt binaries for all platforms plus Docker, and runs fully local — no account, no subscription. 288 commits, 3 releases, MIT licensed. The author's own disclosure: built with Claude Code. curated by the VibeFix editors

CoIsland
大赏
已上线
开发工具claude-code

CoIsland

Your whole engineering stack, centralized in the Mac notch — 35+ connectors, one click hands any alert to your coding agent.

CoIsland lives in your MacBook notch and watches your engineering stack: 35+ connectors for GitHub, Vercel, Snowflake, Jira, Linear, Sentry, PagerDuty, email and calendars. Each monitor runs on its own schedule, and when something fires — a failed deploy, a red CI run, a failed payment — the notch shows the alert with what was found. One click hands it to Claude Code, Codex or Cortex Code, which opens in a terminal right under the alert with the full context; the agent does what is safe, asks you for the rest, and never closes an alert — you do. It is a native Swift/SwiftUI app for Apple Silicon, deliberately serverless: tokens live in the macOS Keychain and the app talks to your tools directly. Two monitors run free forever; a license is a one-time $29.99 early-bird purchase. The maker has run his own days from the notch since July 2026. The author's own disclosure: CoIsland was built with Claude Code (Opus 5.5). curated by the VibeFix editors

proxy429
大赏
已上线
开发工具

proxy429

本地 LLM API 代理:自动吃掉 429 限流和抖动,让 Claude Code / Codex 的长会话不再被上游打断。

proxy429 是一个跑在本地的 LLM API 代理,夹在 Claude Code、Codex CLI 等 AI 编程 agent 和上游模型提供商之间,专门解决一个真实痛点:跑第三方兼容端点时频繁的 429 限流和流式中断。它对 429/5xx 做指数退避自动重试(含藏在 HTTP-200 SSE 流里的限流错误),SSE 保活让客户端在代理重试期间永不超时,还支持模型路由与劫持、OpenAI Responses 与 Anthropic Messages 双协议翻译、加密搜索信封(后续提问直接读上次搜索内容)、以及把显式关思考的请求悄悄升级为 low 思考的细节优化。 配套的本地网页控制台是亮点:实时查看在途流、token 与缓存命中统计、逐请求诊断,请求/响应体按“代理到上游”和“客户端到代理”两侧分开记录——排查 agent 到底发了什么、收到了什么,一目了然。控制台中英双语,跨平台托盘应用单二进制,约 40% 测试代码,MIT 开源。 作者 aaahzb 的原话:这是一个“为 AI agent 造的工具,由 AI agent 造”——他不是 Go 开发者,整个项目是纯 vibe coding 产物。 curated by the VibeFix editors

BagIdea Office
大赏
已上线
开发工具claude-code

BagIdea Office

Your desktop wallpaper becomes a living AI-agent office — real Claude Code sessions that walk, work, delegate and hold meetings.

BagIdea Office turns your desktop wallpaper into a living HD-2D office where every pixel-art character is a real Claude Code session. You play CEO: give an order, a Director agent breaks it down and delegates, and you watch the actual work happen — characters walk to their desks, open files, and run tools in real project folders. A glowing monitor means a live tool call is running; a character walking over to Security means a real permission request. Nothing on screen is canned animation. The office extends itself: agents propose plugins and skills (you approve each one), accumulate shared knowledge in OFFICE.md, and even hold meetings on their own that can turn into written project proposals. It talks over voice, Telegram, Discord and LINE, speaks 14 languages, and lets each agent run a different model brain — Claude, GLM, DeepSeek, Qwen, OpenAI, Gemini, or local Ollama. The project is heavily maintained (241 stars, 98 releases) and MIT licensed. The author's own disclosure: the app was itself built with Claude Code. curated by the VibeFix editors

termist
大赏
已上线
开发工具

termist

Mission control for your coding agents: Claude Code, Codex and OpenCode side by side in one terminal.

termist is a terminal mission control for people who run multiple coding agents at once. Every session becomes a card with a dot: running, waiting for you (a permission or a question), done-unseen, done-seen. Jump straight to the agent that needs an answer, send follow-ups without opening its terminal, resume sessions by name, manage projects as tabs — and the agents keep running when you close the window, owned by a small daemon. The author's disclosure: AI coding tools were used throughout the project — implementation decisions, testing and project direction are the author's. It is at v0.1.4 and moving fast: GitHub PR review (comments, line threads, reviews) already landed, installable on macOS, Linux and Windows. Plus the best-looking terminal UI of the year: Istanbul scenes (Galata Tower at dusk, the Bosphorus) that change with the hour. curated by the VibeFix editors

Cryptid Field Station
大赏
已上线
娱乐

Cryptid Field Station

A vibe-coded cryptid-reporting station: file a sighting, an AI editor scores it, a ranger approves it.

Cryptid Field Station turns hiker sightings into a verified field guide — a Sanity Challenge submission that is genuinely fun to use. File a sighting with story, location and evidence; an AI credibility check plus a real weather cross-check (Open-Meteo pulls the actual sky that night) score the report; then a human ranger approves or rejects it through a real-time triage board — with every transition audit-logged. The workflow even rejects Bigfoot blurs. The author built it entirely by vibe-coding on Sanity: the field guide, a Ranger Dispatch board with live subscriptions, a custom Sanity Studio input that turns trail-cam video URLs into 6-frame GIF strips, and the review process modelled as data (stages, guards, effects). Seven cryptids tracked, twelve sightings filed, eight verified and published. curated by the VibeFix editors

Vibe Games Arcade
大赏
已上线
娱乐claude-code

Vibe Games Arcade

4 款一句话 prompt、单次生成的完整游戏:LOL 峡谷仿真、MC 体素沙盒、PVZ 与杀戮尖塔。

Vibe Games Arcade 是一个 one-shot vibe coding 实验:4 款游戏全部用 Claude Opus 5.5、单次一句话 prompt 直接生成,零后端依赖,纯前端点开即玩。英雄联盟峡谷仿真(精准防御塔仇恨、三路兵线、A* 寻路、完整技能冷却)、我的世界 3D 沙盒(柏林噪声无限地形、体素物理、第一人称漫游)、植物大战僵尸(植物图鉴、阳光经济、波次攻防)、杀戮尖塔肉鸽卡牌(Act 地图、手牌循环、遗物系统、怪物意图预警)。 作者在仓库 README 里明确写道:「本仓库收录的所有游戏作品均为使用 Claude Opus 5.5 进行一句话 Prompt 单次生成 (One-Shot Vibe Coding) 的实验成果。」游戏大厅直接展示了每款游戏的原始 prompt——最短的只是一句话「尽可能利用你的现有能力复刻 Minecraft」。 单次生成做到「能玩的游戏」仍是极限挑战,这个合集就是一份公开的、可亲自验证的能力边界报告。 curated by the VibeFix editors

MechFaber
大赏
已上线
开发工具claude-code

MechFaber

Describe a machine in words — AI picks the parts, writes the CAD and firmware, then simulates it all together.

MechFaber is an AI machine-design engine: you describe the machine you want in plain language, and it searches vendor catalogues for motors, bearings and controllers you can actually order today, draws the CAD around the suppliers' own 3D models, validates fit and load, lays out the full electronics loom, writes the firmware, and then runs the entire machine in simulation — real firmware binary on an emulated MCU (Renode) coupled tick-by-tick with physics (MuJoCo), sharing one battery model. When something is wrong, you reply in a sentence and the whole loop re-runs. The author's own description of the build: Claude Code did all the design work including the firmware. Built over roughly 3 months with a split-agent setup — separate agents for research, calculation, CAD, electronics, firmware and simulation, after a single do-everything agent failed. Clone machines from the library and rebuild them locally, or publish your own. Free while it is early. curated by the VibeFix editors