# What developers are building · 2026-11-06

## AllInOneIME

Mac 输入法，在任何 App 的输入框句首打 @improve、@translate 或 @question 再按回车，就地润色、翻译或提问，选中结果直接上屏，不用复制到聊天窗口；自带 Rime 雾凇拼音，免费，每天 20 次免费 AI，也可以用自己的 Claude / Gemini / DeepSeek key

https://allinoneime.com/zh/

## Pettenderly

在线 AI 宠物写真工具，上传 1–5 张同一只宠物的照片，选择水彩、皇家或影棚风格，生成并下载一幅专属画像

https://pettenderly.com/

## UsageRail

API 计费与限流工具，几行中间件给自己的 API 发 Key、按套餐限流和配额，用量通过你自己的 Stripe 按量扣费；Key 和限额在你的服务进程里校验，请求不经过第三方；每月经它计费 1000 美元以内免费 - 文档

https://usagerail.com/

## Parallel Parking Simulator

侧方停车模拟器（免费），打开网页即可练习打方向、倒车入位和回正，附分步停车指南

https://parallelparkingsimulator.com/

## Murmury：视频会议时，根据你自己的文档实时给出简短回答（Mac，可完全本地运行）

The open-source Cluely alternative for Mac: live answers from your own docs during video calls, hidden from screen sharing. Local Whisper, your own AI key.

https://github.com/GianlucaBave/murmury

## iLoveDevTool：免费在线开发者工具箱（40+ 工具，代码运行 / 格式转换 / 文本处理）

自荐一个开发者工具站：iLoveDevTool 40+ 个免费在线开发者工具的集合，无需注册： 代码运行：Python、PHP、Java、C++、Go 在线 playground，支持多文件项目与分享链接 数据转换：JSON 转 Excel / CSV / YAML / PHP / Python，XML 校验（DTD / XSD），Excel、CSV 转 JSON 文本处理：文本对比（diff）、正则测试、大小写转换、Markdown 预览 开发常备：Unix 时间戳转换、时间换算、Base64 / URL 编解码、JWT 解析、哈希生成、二维码生成 界面支持 20 种语言（含简体 / 繁体中文）和暗色模式。全站免费，无付费档。 利益相关：我是站长，本帖为自荐。

https://www.ilovedevtool.com/zh-CN

## CodeYear：本地优先的年度代码报告生成器，直接读取本地 git、不上传代码

本地优先的年度代码报告：读取本机 git，生成中英双语 HTML、海报、长图和 GitHub 主页卡片。 | Local-first annual coding report from your own git history.

https://github.com/Qiongkura/codeyear

## Flipic：给摄影师挑片用的开源 Mac 看图软件

Flipic: free, open-source image viewer for Mac. Double-click a photo and page through the folder with the arrow keys.

https://github.com/xmasdong/flipic

## autorekt：让 AI agent 读懂全球市场的命令行数据工具

A CLI that gives AI agents market data from free public sources: prices, dollar liquidity, swap spreads, positioning, options gamma, prediction-market odds and news, as compact text.

https://github.com/autorekt/autorekt

## Agentic Kingdoms：真人玩家与 AI 智能体同场的网页策略 MMO

Agentic Kingdoms 是一款免费的中世纪王国战争 MMO，手机和电脑浏览器打开即玩，无需下载，可以用游客身份直接开始。 聊天在 11 种语言之间实时翻译，不同国家的玩家可以各用自己的语言交流。 真人玩家和 AI 智能体在同一个持续运转的世界里游玩，操作频率限制相同（智能体每天攻击和侦察同一名真人玩家的次数另有上限）；智能体名字旁带有所有玩家可见的 AI 标签。 智能体通过公开的 MCP 服务器或 HTTP/WebSocket API 加入，无需密钥；接入说明是一份公开文档： 试玩：

https://agentickingdoms.com/skill.zh.md

## Agent Ping — 菜单栏上告诉你哪个 AI 编程 agent 在等你的开源工具

Menu-bar panel that tells you which AI coding-agent session needs you right now. macOS + Linux, reads only the logs your tools already write, makes no network calls.

https://github.com/DreamOfXM/agent-ping

## Custom Song Gift：用真实回忆与名字，在几分钟内谱写专属定制音乐

推荐产品：Custom Song Gift Custom Song Gift 是一款基于故事与情感驱动的 AI 专属歌曲定制工坊。用户只需输入纪念日细节、专属昵称、感人瞬间或生活琐事，即可在数分钟内将真实回忆谱写为富有感染力的定制歌曲。 核心特点与解决的痛点 回忆语义锚定与结构化歌词：不同于泛泛的 AI 音乐提示词，系统提取人名、专属故事与关键情感锚点，编排成严谨押韵的主歌与副歌，避免歌词泛化与空洞。 16 种录音室级曲风：涵盖温暖木吉他（Acoustic）、流行钢琴抒情（Ballad）、乡村叙事（Country）、流畅 R&B、Lo-Fi 等，具备丰富的乐器编曲质感。 专属数字礼品播放页：每首歌曲交付一个可永久分享的黑胶唱片机在线播放页，支持双语同步滚动歌词、唱片封面定制与无损 MP3 音频下载。 平价亲民与免订阅：相比传统定制歌曲服务动辄 100~200 美元的高昂门槛与数周交付周期，.90 起即可完成一首录音室级完整歌曲，且无需月度循环订阅。 Agent 与工作流生态接入：提供公开的 Model Context Protocol (MCP) 端点（ n8n 工作流模板，便于开发者与智能体调度。 官网体验：

https://customsonggift.org/

## 卖家图片策划 Free：把产品资料整理成三张有证据边界的拍摄任务卡

卖家图片策划 Free：商品资料 → 最多三张 FABE、文案与拍摄任务卡。Agent Skill for e-commerce product-image briefs; plans, not photos. 自定义免费许可，非开源。

https://github.com/PDBen-Auto/seller-image-brief-free

## HeadlineArena：用真实事件检验 AI Agent 的预测判断能力

大家好，我们在做 HeadlineArena，一个面向 AI Agent 的真实世界预测评测平台。本帖为项目方自荐。 网站介绍 我们围绕结果尚未揭晓的真实事件发布预测任务，让 Agent 在截止时间前提交预测和判断依据；结果公布后，按任务规则核验、评分，并形成可持续查看的表现记录。 目前网站可以查看预测任务、已结算结果和公开排行榜，也提供 Agent 接入指引。任务涉及能源、粮食、利率和官方经济数据等。目前公开结果评估的是包含模型、工具和提示词的 Agent 系统。 我们的服务 对企业AI团队：为做大模型后训练、Agent 开发和评测提供前瞻预测评测。我们也想请教后训练评测的同学，这类真实世界反馈，能否帮助发现模型的判断偏差，并为后续优化提供参考？欢迎体验、接入自己的 Agent，也欢迎讨论任务设计和评分方法。 对开发者：如果你对预测感性，可以直接接入你的agent在平台做预测赚取积分，目前积分可兑换token，未来还有更多兑现的手段。 对标产品 我们对标的国外产品有： forecastbench futuresearch 更多信息

https://headlinearena.com/

## 定时自动网页签到助手

浏览器扩展：一键完成所有网站的每日签到

https://github.com/qinruimac/checkin-extension

## CodeX 重度患者必备！原生插件看板管理并发对话，组织对话合集

A Kanban board for Codex power users to manage concurrent chats, tasks, and usage history. 专为 Codex 重度使用者设计，看板模式管理并发对话，组织对话合集

https://github.com/jingchengyin/codex-kanban-plugin

## 蜗牛快搜 — Windows 中文桌面文件检索工具

Windows 本地文件即时搜索工具:全盘索引、输入即搜,通配符/正则/SQL/Lua 五种搜索模式,多窗口、插件系统、6 种界面语言,纯 Direct2D 自绘 UI。

https://github.com/Winlpl/SnailQuickSearch

## 小维助手：在 Android 手机上运行 Pi / DeepSeek Harness 的 AI Agent

Open-source AI assistant that operates your Android phone for you, built on Pi · 小维助手

https://github.com/zjw1918/WeiAssist

## MiniToolDesk: 100% 浏览器本地运算、零数据上传的极客开发者工具箱

软件名称 MiniToolDesk () 核心特点与解决痛点 日常开发中，工程师经常需要使用 JSON 格式化、代码/文本比对、Hash 计算等小工具，但市面上大量在线工具广告泛滥，甚至将敏感数据上传到第三方后端，存在极大的数据泄露隐患。 MiniToolDesk 采用严格的“100% 浏览器本地计算”架构： 1. 零服务端数据交互：利用 Web Crypto API、HTML5 Canvas 和 Web Workers，所有计算与转换均在本地内存中完成，断网可正常使用。 2. 多语言矩阵与侧边栏支持：支持中文、英文、日文等 7 种语言，支持 PWA 离线运行，并提供专属 Chrome/Edge 侧边栏扩展（Alt+M 随手呼出，无需频繁切屏）。 3. 内置 AI 数据脱敏与 AST 矩阵：提供专用的本地 Prompt 脱敏工具（向 ChatGPT 提问前一键屏蔽 API Key 与敏感数据），以及 JSON 一键转 Zod / Go Struct / Pydantic 等类型定义。 4. 全套免费无广告：界面参考 Raycast / Linear 设计规范，零弹窗、零强制注册。 欢迎体验与交流：

https://minitooldesk.com/

## 经典电影台词，学英语， WEB版开源， docker一键部署，即开即用！

用照片学英语。无需登录的开源 Web 学习器，支持 PicLex 卡组。

https://github.com/zhaoolee/piclex-web

## Fun - 基于 C 转译且内置 M:N 虚拟线程的静态类型系统级编程语言

Fun is a statically-typed language that transpiles to C, combining safety and performance with C's efficiency.

https://github.com/omdxp/fun

## AgentCraft：64 条 GPT Image 2.6 提示词，每条附成图和出处

OpenAI 11 月 4 日发布了 GPT Image 2.6，ChatGPT 里叫 Images 2.6，新增 GIF 动图。这一页收了 64 条它的提示词，每条原样摘自 OpenAI 官方的 Image prompting guide 或创作者在 X 上的帖子，旁边放着生成的成图或 GIF，并链到出处，点开就能复制全文。 页首还整理了这次的变化：API 里 flare 和 sunburst 两个同价模型的区别、新增的圈选局部改图和 SVG 导出、每百万 token 的价格。 免费浏览和复制，不用注册，中英双语（英文版 ）。 披露：我参与维护 AgentCraft，这是自荐。

https://agentcraft.dev/zh/prompts/image/gpt-image-2-6/

## BoopFace：AI Bot 表情组件，零依赖SVG

BoopFace (formerly Agent Robot Avatar): a lightweight, expressive robot avatar component for AI bots, agents and interactive apps.

https://github.com/CX-ArtLab/BoopFace

## AI Focus — 同时开很多个 AI 时用的专注力面板（Windows）

同时开着网页版 ChatGPT / Claude、几个 Claude Code 终端、Codex 桌面版，窗口一多就不知道谁跑完了、谁在等你。AI Focus 把所有会话集中到一个面板：谁在等你、等了多久、最后回复了什么，一眼看清；Alt+W 跳到等你最久的会话，Alt+E 跳回来，能定位到终端的具体标签页。数据全在本机，不联网、不用账号。仅 Windows，浏览器扩展需手动加载。

https://github.com/jy1989/AiFocus

## [Tool Recommendations] 4 Highly Recommended Free AI Tools You Can Use Without Signing Up

PixEditor AI is a free AI photo editor that lets you edit and transform photos online using simple text prompts. Easily replace backgrounds, remove unwanted objects, restore photos, and adjust lighting or colors with AI. Customize image styles and visual effects to create unique visuals without professional editing skills. No registration or login is required, and you can edit images for free without watermarks. It's ideal for eCommerce product images, social media content, and everyday photo editing. Key Features: AI-powered online image editing that lets you modify photos quickly using text prompts. Supports background replacement, object removal, photo restoration, and lighting and color adjustments. Customize image styles, backgrounds, and visual effects to transform photos creatively. Free to use online with no registration or login required and no watermarks. Ideal for eCommerce product image optimization, social media graphics, and everyday photo editing. 2. AI Floor Plan Generator – Floor Plan Maker Tool: Floor Plan Maker – Free Floor Plan Maker AI

https://pixeditor.ai/

## paper-trends：把学术会议十年的论文变成可交互趋势报告

Publication-trend analysis for any academic conference, powered by OpenAlex

https://github.com/Minjia-Hu/paper-trends

## Funnybit.cn妙比特：Vibe Coding 作品与 AI 应用灵感发现平台，汇集独立开发者创作的 AI 应用和实际作品案例

推荐一个最近在做的小项目：妙比特 funnybit。 网站地址： 现在用 AI 写代码越来越方便，但一个常见的问题是：打开 Codex、Claude Code 或者DSH后，反而不知道做啥，拿着锤子找钉子。 所以我做了 funnybit，希望把大家用 Vibe Coding 创作的有趣作品集中展示出来，让更多人获得灵感。 目前主要有两个栏目： 1. 发现作品 收集和展示 Vibe Coding、AI 应用开发相关的实际作品，包括网站、小工具、互动应用等。可以按分类浏览、查看作品介绍，并通过原作品链接进一步体验。 除了发现其他开发者的作品，我们也在尝试自己用 AI 构建一些有趣、实用的小应用。 2. AI 头条 整理 AI 产品、模型和开发工具的最新动态，同时提供原创 AI 日报和周报，希望让关注 AI 创作的人更方便地了解新变化。 目前网站已经上线，仍在持续迭代。 我希望 funnybit 不只是一个网址收藏夹，而是一个能让人产生「这个创意不错，我也想动手试试」的 AI 作品灵感广场。 欢迎大家体验，也欢迎反馈建议！ 在线访问：

https://funnybit.cn/

## ManaWG：可自托管的合租室友应用（账单分摊、打扫轮值、购物清单、聊天）

Free, open-source app for shared flats and roommates: split expenses, cleaning rota, shopping list, calendar and chat. A self-hostable Splitwise alternative with Docker.

https://github.com/Abuarchiv/manawg

## ACP AI Research：中英双语 AI 日报，每篇附原始来源和日期

一个个人做的 AI 日报站，中英双语，每天北京时间 10 点更新，最多 5 篇。 分为 AI 新闻、工具评测、研究解读、免费资源、提示词工作流五类 每篇开头一句话结论，文末列出原始来源及发布日期，方便自己核实 支持 RSS（ Telegram 频道 @acprelation 说明：内容由 AI 根据公开来源整理，经自动质量评分筛选后发布，站内关于页有方法说明。技术栈：Astro + Cloudflare Workers + D1。 谢谢阮老师的周刊！

https://acp.dpdns.org/zh

## Mini Timer：免费、纯净的 PPT 悬浮计时助手

Free, offline floating timer for PowerPoint presentations. Windows & macOS. 免费、纯净的 PPT 悬浮计时助手。

https://github.com/michael-chow-arch/mini-timer

## KinetAios:Claude Code 与 Codex 双引擎桌面仪表盘(多会话同窗管理)

Local-first multi-engine AI agent dashboard — Claude Code, Codex & built-in ReAct in one window. MCP support, skills auto-loading, long-term memory. No account, no relay.

https://github.com/phinn/KinetAios

## florilegium

Procedurally generated Delft-blue floral page margins for LuaLaTeX, with plants and visitors sampled from genomes and a TikZ library that also draws them under pdfLaTeX

https://github.com/kevinraymond/florilegium

## Disc Image Studio

这个工具可以将指定的图片刻录到光碟上，使得图案在光碟表面上肉眼可见。

https://github.com/JiaFeiMiao-K-Cat/DiscImageStudio

## OutCrap

一个免费的手机 App（苹果和安卓），可以根据照片分析地质地貌，也可以根据 GPS 定位给出地质信息。

https://outcrop.rocks/

## SOML

A config format for humans

https://github.com/soml-lang/soml

## Casnode

Open-source forum with the look and features of Discourse, built with Go, Next.js and Casdoor SSO. Works with MySQL, PostgreSQL, SQLite and SQL Server

https://github.com/casnode/casnode

## 澜姿

澜姿：从姿势线稿和衣装集给社区鲸鱼娘大肥鱼模特出图，调整姿势加变装，也可装进 DeepSeek Harness。Lanpose generates images of the community whale-girl Dafeiyu from pose sketches and costumes, and installs into DeepSeek Harness.

https://github.com/wzc520pyfm/gen-ai-pose-image

## Crossbow

Proxy Client

https://github.com/Richard-Zhang1019/Crossbow

## WeRead Book Exporter

📚 微信读书无损全书导出工具 — 一键导出双字体精排 PDF / EPUB / 纯 Markdown 笔记（适配 Obsidian & Notion），攻克网页端翻页漏字问题。

https://github.com/WANGXIN-567/weread-book-exporter

## Linkit

AI bookmark manager for desktop — local-first, offline, read-it-later and PKM tool with AI summaries, semantic search, knowledge graph, and free self-hosted sync. 100% free, no paywalls, no ads.

https://github.com/blue-idea/linkit

## halfsize

Compress any video to an exact size (Discord or WeChat 20 MB, WhatsApp 16 MB) locally with FFmpeg, then prove the quality with VMAF/SSIM. No upload, no account.

https://github.com/YKLovePython/halfsize

## HackerTube

一个 Youtube 视频网站，汇集了“黑客新闻”论坛上提到的所有 Youtube 视频，可以用来寻找值得看的科技视频。

https://hackertube.org/

## 开始学习图论

一本面向初学者的图论入门教材（英文），免费阅读和下载。

https://vertex.degree/

## Python 数据科学手册

该书向初学者介绍用于数据处理的一些 Python 基本工具（比如 NumPy、Pandas），出版于2016年，免费阅读。

https://jakevdp.github.io/PythonDataScienceHandbook/

## Varin：全能型 Super-Agent Harness —— 细致打磨的 UI/UX、丝滑 Chat 交互与全栈工具生态

The open-source agent harness, ready out of the box and customizable in depth. Tools, context, durable execution, permissions and recovery for code, research and office work, with a long-running super-agent Bot that leads a team of agents.

https://github.com/Youzini-afk/Varin

## lmux：为 AI 编程 agent 打造的轻量 macOS 终端

A light, native macOS terminal for living in Claude Code and Codex all day.

https://github.com/sooftware/lmux

## 一行安装的中文求职 Agent Skills 套件与 12 仓公开横评（支持可复算简历匹配 MCP）

一行安装的中文求职 Agent Skills 套件（岗位搜索/简历匹配/面试陪练/查公司）+《中文 AI 求职 Skill 公开横评》数据。Claude Code·Codex·MCP

https://github.com/ouxxyy/ouba-classmate

## 基于 Excalidraw 的自托管团队白板，支持单点登录和实时协作

Your team's whiteboard, on your server. Self-hosted Excalidraw with accounts, folders, and search.

https://github.com/maciekzygmunt/excaliself

## zed-industries / zed

Code at the speed of thought – Zed is a high-performance, multiplayer code editor from the creators of Atom and Tree-sitter.

https://github.com/zed-industries/zed

## rails / rails

Ruby on Rails

https://github.com/rails/rails

## mpv-player / mpv

🎥 Command line media player

https://github.com/mpv-player/mpv

## rust-lang / rust

Empowering everyone to build reliable and efficient software.

https://github.com/rust-lang/rust

## react / react

The library for web and native user interfaces.

https://github.com/react/react

## angular / angular

Deliver web apps with confidence 🚀

https://github.com/angular/angular

## symfony / symfony

The Symfony PHP framework

https://github.com/symfony/symfony

## prometheus / prometheus

The Prometheus monitoring system and time series database.

https://github.com/prometheus/prometheus

## microsoft / vscode

Visual Studio Code

https://github.com/microsoft/vscode

## nushell / nushell

A new type of shell

https://github.com/nushell/nushell

## etcd-io / etcd

Distributed reliable key-value store for the most critical data of a distributed system

https://github.com/etcd-io/etcd

## openssl / openssl

General purpose TLS and crypto library

https://github.com/openssl/openssl

## cloudflare / security-audit-skill

A coding-agent skill for multi-phase security audits with independently verified, machine-readable findings

https://github.com/cloudflare/security-audit-skill

## astral-sh / ruff

An extremely fast Python linter and code formatter, written in Rust.

https://github.com/astral-sh/ruff

## halo-dev / halo

Halo 是一款强大易用的开源建站工具，从个人博客、知识库，到企业官网、在线商城，Halo 都能助您轻松实现，一站式满足您的多样化建站需求。

https://github.com/halo-dev/halo

## gaotianliuyun / gao

FongMi影视和tvbox配置文件，如果喜欢，请Fork自用。使用前请仔细阅读仓库说明，一旦使用将被视为你已了解。

https://github.com/gaotianliuyun/gao

## soybeanjs / soybean-admin

A clean, elegant, beautiful and powerful admin template, based on Vue3, Vite7, TypeScript, Pinia, NaiveUI and UnoCSS. 一个清新优雅、高颜值且功能强大的后台管理模板，基于最新的前端技术栈，包括 Vue3, Vite8, TypeScript, Pinia, NaiveUI 和 UnoCSS。

https://github.com/soybeanjs/soybean-admin

## lionsoul2014 / ip2region

Ip2region is an offline IP-to-Region localization library and IP data management framework with both IPv4 and IPv6 supports, 10-microsecond level query efficiency, xdb search client for many programming languages

https://github.com/lionsoul2014/ip2region

## langbot-app / LangBot

Production-grade platform for building agentic IM bots - 生产级多平台智能机器人开发平台/ Agent、知识库编排、插件系统 / Bots for Discord / Slack / LINE / Telegram / WeChat(企业微信, 企微智能机器人, 公众号) / 飞书 / 钉钉 / QQ / Matrix e.g. Integrated with ChatGPT(GPT), DeepSeek, Dify, n8n, Langflow, Coze, Claude, Gemini, GLM, Ollama, SiliconFlow, Kimi, openclaw / hermes agent, deerflow

https://github.com/langbot-app/LangBot

## higress-group / higress

🤖 AI Gateway | AI Native API Gateway

https://github.com/higress-group/higress

## MaaAssistantArknights / MaaAssistantArknights

《明日方舟》小助手，全日常一键长草！| A one-click tool for the daily tasks of Arknights, supporting all clients.

https://github.com/MaaAssistantArknights/MaaAssistantArknights

## Anduin2017 / HowToCook

Programmer's guide about how to cook at home.

https://github.com/Anduin2017/HowToCook

## notionnext-org / NotionNext

Turn your Notion workspace into a fast, customizable website. Built with Next.js + Notion API, with multi-platform deployment and no self-hosted server required.

https://github.com/notionnext-org/NotionNext

## yangzongzhuan / RuoYi-Vue3

🎉 (RuoYi)官方仓库 基于SpringBoot，Spring Security，JWT，Vue3 & Vite、Element Plus 的前后端分离权限管理系统

https://github.com/yangzongzhuan/RuoYi-Vue3

## 521xueweihan / HelloGitHub

分享 GitHub 上有趣、入门级的开源项目。Share interesting, entry-level open source projects on GitHub.

https://github.com/521xueweihan/HelloGitHub

## pymumu / smartdns

A local DNS server to obtain the fastest website IP for the best Internet experience, support DoT, DoH, DoQ. 一个本地DNS服务器，获取最快的网站IP，获得最佳上网体验，支持DoH，DoT，DoQ。

https://github.com/pymumu/smartdns

## Cursor Craft

Custom cursors for Mac

https://www.producthunt.com/r/BCO35XHIOIH5EM?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Meadow

Hidden songs on your walk, chosen by people, not algorithms

https://www.producthunt.com/r/KD2FP5YECF3KLS?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Native Tasks for Obsidian

A native Mac app for your Obsidian tasks

https://www.producthunt.com/r/XF4DEVS6KJHUGU?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## pwned.coffee

A live board of security flaws

https://www.producthunt.com/r/L5JM67PP76ZJW6?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## PostMyPace

Turn your run into a ready-to-post Instagram Story or Reel

https://www.producthunt.com/r/ATHUSO6VSZSN5R?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Openroster

Which "Open" companies are actually open source

https://www.producthunt.com/r/WJVOKB2CCZDDU2?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Meshia

Your autonomous research lab

https://www.producthunt.com/r/5CMNIPUNTQFX7X?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Chronos DB

The database for the autonomous AI era.

https://www.producthunt.com/r/2RUZVT2TUIFYVF?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Textwhisk

Turn selected text into quick actions on your Mac

https://www.producthunt.com/r/FQT3F5CBOLRK5R?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## axel.md

Agent orchestration in the terminal, designed for Rex

https://www.producthunt.com/r/7QBNO4SPNFWPBT?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## SafeGrd

Safae database backups that AI can restore but not delete

https://www.producthunt.com/r/WHRUU5UCWAR7QG?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Releasebeam

Release infrastructure for developers

https://www.producthunt.com/r/T22UL5MBNS4M4K?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Alysis Code

Open-source AI coding agent

https://www.producthunt.com/r/ZJZGOZVSMYBDXG?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Translucent AI

Test infrastructure for hardware teams

https://www.producthunt.com/r/T5U3FBUHHDAIJE?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Shipnoted

Changelog + feedback board for closing the loop on your git

https://www.producthunt.com/r/KFQDRX4FWXZITG?utm_campaign=producthunt-api&utm_medium=api-v2&utm_source=Application%3A+community-pulse+%28ID%3A+298617%29

## Agentic Identity Broker, OAuth Delegation for AI Agents

Identity broker for AI agents with OAuth2 delegation and consent management

https://github.com/zalando-incubator/agentic-identity-broker

## An open source webbluetooth and NFC browser for iOS

An open-source iOS browser for developers. Test JavaScript web apps on iPhone with WebBluetooth (BLE/GATT) and WebNFC (NDEF tags).

https://github.com/hito-dev/wand-browser

## HackerRepos – HackerTube for repos mentioned on HN

Karpathy posted hackertube.org 5 days ago, and mentioned that it took him 2 hours to build. A day after another fellow posted hackerbooks.org https://news.ycombinator.com/item?id=50326821 , and made it to the top of HN. Some hours after someone else made it to the top of HN with the same for wikis: https://wikihack.netlify.app/ I found particularly interesting the "Trends" section on https://hackerbooks.org/ That night, I set out to build something similar but for Repos. So I built https://HackerRepos.fyi . There are 57,316 repositories and 667,508 mentions indexed. I originally made three main sections: Top of all time, Most Recommended, and Trends - per HackerBooks inspiration. It looked/felt good, so I wanted to take it further. I am a big fan of https://github.com/trending , and frequent it often, looking for inspiration. So I built the "Trending now section" https://hackerrepos.fyi/#/trending section. Trending now gives in a quick colour-coded sentiment snapshot of the repos trending on HN, using a grid/map in the style of https://finviz.com/ . When you click on a repo, it gives you a summary of the chatter, a colour-coder tag (Favorable, Critical, Debated, Informational), a graph of the mentions over time and the comments below the graph. I also added a section on every called "Where it’s used", to showcase real world use cases. I realized populating this section could be costly, so I placed it behind a paywall for those that care for it. Pro users can click "Find real-world uses", and the app goes and finds a real-world use case. I am working on a few other features I want to ship, but they are not ready yet. For example, I want to build a reverse product-to-stack exploration feature, where a user could enter or describe a product, and the app could provide a set of stacks to fit it, and perhaps, supporting comments. Also, I want to build task-based comparisons. I have a task, and I am debating between some repos to use. I could describe the task and enter the repos I have in mind to solve for the task, and have the app suggest the better one. But I am glad I stopped and I am posting this now, because it has been a long time since I have finally managed to actually ship something from start to finish. I have a folder full of projects I started and did not ship. As I was struggling with this yesterday, whether I should make this public now or keep fiddling with it, I went to Mass and the gospel was Luke 14:25-33. The following part struck me: "And indeed, which of you here, intending to build a tower, would not first sit down and work out the cost to see if he had enough to complete it? Otherwise, if he laid the foundation and then found himself unable to finish the work, anyone who saw it would start making fun of him and saying, Here is someone who started to build and was unable to finish." So here it is. I finished this little tower. I hope you enjoy it, find it helpful, and if you care for it, sponsor it.

https://hackerrepos.fyi/

## Inference business simulator – don't go broke, rent GPUs, serve tokens

A browser game about running an inference business for one year. You rent GPUs, choose which open-weights models to serve on which of them, and give spare GPUs to Elise to find speedups for the engine.

https://makermaker.ai/breakthroughs/inference/simulation/

## A C++ Compiler That Is Faster Than GCC and Clang

I've been working on a scaffold to teach an agent how to build a self-hosting C++ Compiler from scratch for the past 6 months. One of my recent agent-runs recently achieved self-hosting for less than $500 in API tokens, but the most interesting part is that the reference implementation (which takes the best ideas from every run) can now compile C++ code 4x faster than GCC and Clang.

https://github.com/vishvananda/cppgm-extended

## HackerGit – Every GitHub repo shared on HN

I love the trend karpathy started, I hope this helps add to it :)

https://hackergit.org/

## Ueddit.com – Reddit Without Logging In

Example: https://www.reddit.com/r/Steam => https://www.ueddit.com/r/Steam => Get redirected to a working RedLib instance. Simple replace the "r" with an "u" (reddit => ueddit)!

https://ueddit.com/

## Cabinet of Distilled Pixels

I got this idea from the recent "Sparks of Weight-Space Programming" post. I thought the abstracted game mechanics looked quite nice, and wanted to expand on them a bit.

https://cjauvin.github.io/pixel-cabinet/

## Rindraw – drawing and animation app for iPad, iPhone and Mac

Rindraw is a drawing and painting app for iPad, iPhone and Mac. Every line you draw stays a line: recolor it, bend it or scale it up long after, and it stays sharp.

https://rindraw.com/

## Overhead – the planes overhead, painted on an e-ink frame

The planes flying over your house, each painted from a photograph of that exact aircraft, on an e-paper frame. Four widgets for Tesserae.

https://github.com/dmellok/tesserae-widget-overhead

## BeepGrid – Lightweight open source agent comms

Fast communication infrastructure for agents and the people they work with.

https://github.com/beepgrid/beepgrid

## Ski Voice-First Multilingual Neutral Coding Harness

Glad to showcase SKI ( https://justski.it ) , a voice-first coding harness. It supports multiple agents from different providers, just like conductor used to. You can run claude code, codex, antigravity etc. You can also add multiple subscriptions of Claude code or others, which will automatically switch when the limits are hit so that you can continue working on projects seamlessly Voice first - You can talk to it, and it will talk back and keep you updated on tasks and status. You could also ask it to speak only when needed and it will adapt. (It doesn't read out the code line by line, unless you specifically ask it to). It is multilingual and supports 80+ languages. It can automatically convert the audio in your language to English for the models to work on. There is a optional review window if you need to verify it before sending to the agents. Chain reactions: Maybe you need one agent to do a task, and then hand it over to the next agent. We call it chain. And you can just do it using voice easily in SKI. Lives in notch (can be hidden behind notch as well, but also has a bigger UI like conductor) Different voice for different sessions: Distinguish projects by voice. Can take if you ask it to (will show flash) One agent for all - Experimental feature where a single agent manages multiple sessions by voice, routing your requests on its own and keeping track of tasks across different sessions. UI pop ups for things that need your attention - When in notch, agents will pop up options and explain what each options are, when they need an input from you. For instance, if there are 3 options for some task, the UI will show the 3 tasks as a floating window, which you can either click or choose by voice. Agents will also explain by voice if needed. Floating taskbar like a widget - Floats when enabled, showing all tasks in different sessions and different agents, along with current progress and step. Can give voice inputs by just clicking on them and talking. Loops - available for all agents as a harness tool Computer use - Stays in notch and does without interfering your work. Git integrated: Can create sessions from PR, merge, see comments etc. All by voice, text and UI buttons. Also custom branch names with configurable prefix. Pricing: Text usage - Unlimited, free forever, No limits Voice: Free 300 credits per month.

https://justski.it/

## The web needs more creativity and whimsy

Someone commented that all UIs are starting to look very much the same and very AI lately. I think we play and experiment more.

https://ignition.github.io/press/

## An aquarium game with a real nitrogen cycle

Keep a virtual aquarium with a real nitrogen cycle: cycle the tank, feed, change water and stock fish without losing any.

https://fishtankatlas.com/play/aquarium/

## Grid – A Hold'em arena where your AI agent plays your strategy (Beta)

Hi HN, my friend's team and I have been building a game called GRID, a dog poker game with an AI twist. Under the cute dog graphics, GRID is an arena where LLMs compete, with no-limit hold’em as the medium. You duke it out as a strategist, vibe coder, and bot maker. Why: models get compared on benchmarks all the time, but rarely as my agent against yours in its truest form. GRID is our attempt at creating that arena, because we believe this type of game will exist in the future, as AI becomes as common a tool as the browser. How it works: * Paste a prompt and an MCP config into Claude Code, Codex or any MCP client. The agent waits for its turn and sends actions back through tool calls. * The strategy is text written in the game, up to 6,000 characters. We thought it’d be cool to have a preset, “Play like Elon”, to show you what you can do. * You can change the strategy any time, and it applies from the next action. You can even ask your AI to change it for you, based on what it learns about its opponents’ strategies from the match history. * Free beta. Play chips only. We just opened up the game today and are gathering data on how to steer this ship next. Rough edges: * MCP can be finicky, especially if you refresh your browser while in a match. If this happens and you can’t get your agent to connect, wait 5 min and retry | log out using the settings button at the top left and log back in. Stability is currently our top priority. * MCP also uses a lot of tokens, more than the game should, because every tool call costs tokens. We plan to overhaul this with a new connection type in a future version so it’s much more efficient. * There’s a built-in local model for folks who don’t want to use their own AI. It plays basic hold’em OK but struggles to follow written strategies. If you want to try it, I’d recommend leaving your strategy blank. Sadly, it plays best with no instructions. * House bots fill empty seats. We made this choice so matches happen even when only a few people are online. Honestly, I’m most curious how you’d write a strategy that holds up against everyone else’s agents. And if MCP setup breaks on you, please tell me. We'll be in the thread as long as possible.

https://penguinbase.com/game-install?game=192471728&ref=hn

## 做了一个开源的 YouTube 桌面客户端 Lumtub：原生播放，推荐在本地跑，支持 Windows / Linux / macOS

Lightweight Flutter Windows YouTube player using media_kit, libmpv, and yt-dlp

https://github.com/0x696c757a696f/lumtub

## 做了一个按月整理相册的 iPhone App：滑动取舍照片，也能先看大视频

大家好，我是 Photo Cleaner: Swipe Sorter 的开发者。想在这里介绍一下这个相册整理工具，也听听大家平时怎么处理越积越多的照片。 它的主要流程是先选一个月份，再逐张滑动保留照片或标记删除，把整理范围缩小到一个月。遇到相似照片，可以放在一起比较；视频可以按文件大小排序，先检查占空间较大的文件，再决定是否批量删除。最终删哪些内容由你选择。 我想了解的是：按月分批整理是否更容易坚持？还是你更希望优先处理相似照片和大视频？如果试用时遇到不顺手的地方，也欢迎直接指出。 支持 iPhone / iPad ，需要 iOS / iPadOS 26 或更高版本。免费下载，包含付费内购，美区价格为每月 $0.99 、每年 $3.99 ，或 $6.99 永久解锁；不是所有功能都免费。其他地区的价格及可下载情况以对应 App Store 为准。 美区 App Store： 文案由 AI 辅助整理。

https://www.v2ex.com/t/1255115

## 做了个 Claude Code CLI 侧边栏

Plugins for AI coding agents

https://github.com/itaober/agent-plugins

## 做了个 Agent 版的「Adobe + Office 全家桶」，能在 codex、claude code 中直接使用

大家好，最近做了一个在线办公和创作工具站 EasyRunCraft ，分享给大家试试。 可以把它理解成一套类似「 Adobe + Office 全家桶」的工具：文档、表格、PPT 、PDF 、图片、视频、音频、CAD ，都有对应的编辑器。打开浏览器就能用，也可以接到自己的 Agent ，让它帮忙处理文件。 先说一下来源：这 13 个编辑器来自 ArtCraft 的开源 Crafting Apps 项目，编辑器本身的能力由原作者和贡献者完成。 我主要做的是在线整合、统一账户，以及内置 AI 和 MCP 接入。感谢原作者和贡献者，原项目源码在这里：GitHub 。 用起来有几种方式： 平时可以直接当在线软件用，打开文件、修改、保存，不需要安装。临时改个 PDF 、处理图片、调整表格，都可以找到对应的编辑器。 编辑器里也加了 AI 助手，可以让它帮忙修改文档文字、更新表格、调整图片图层。改完直接看结果，需要时可以撤销。具体支持哪些操作，各个编辑器会有区别。 如果你已经在用 Codex 、Claude Code 等 Agent ，也可以通过 MCP 接进来，在原来的对话里让它处理文件，再拿到修改后的副本。接入说明放在了这里：MCP 。 我想做的就是给 Agent 配上一套办公和创作工具，让它能直接帮你把文件改好。 手动编辑、保存和导出免费，内置 AI 和 MCP 操作使用积分，各个工具共用一个账户和余额。 欢迎大家试用、提意见。也想听听，你平时最想交给 Agent 处理的文件操作是什么？

https://www.v2ex.com/t/1255104

## vibe 了一个压缩图片到指定大小的工具

有些网站上传图片会限制大小，这个工具可以压缩图片到指定的大小，例如: 10mb -> 500kb 这也是一个 npm 包可以集成到你的项目上传组件中,省去了让用户手动调整的麻烦： ```javascript import imgcap from 'imgcap' // Social media avatar (500KB limit) const avatar = await imgcap(file, { targetSize: 500 * 1024 }) // With format conversion const webp = await imgcap(imageFile, { targetSize: 300 * 1024, outputType: 'image/webp' }) ```

https://www.v2ex.com/t/1255100

## 做了个 V2EX 看图插件，截图里的小字终于不用新开标签页看了

逛 V2EX 时经常碰到一些截图，缩在正文或回复里，字太小看不清。想看细节就得打开原图，再切回帖子，长图看起来也不太方便。 于是做了个 Chrome 小插件 V2Peek - V2EX 图片放大镜，主要解决这个问题：不用离开帖子，就能放大图片、查看细节，看完继续读。 用法很简单： - 点击正文或回复中的图片，直接弹出大图。 - 滚轮放大、缩小，以鼠标位置为中心查看细节。 - 按住左键拖拽，移动到想看的区域。 - 一键适应窗口，按 Esc 关闭。 - 图片外层链接指向原图时，直接打开查看器，避免跳转；其他链接保留原有行为。 目前免费，无广告，不用注册，不收集个人数据。 安装地址： 安装后刷新 V2EX 页面即可使用。独立开发，与 V2EX 官方无关联。 功能不多，就是想把看图这件小事做顺手一点。欢迎试用，有问题可以贴具体页面和复现步骤。

https://www.v2ex.com/t/1255096

## 做了个 Chrome 插件 LookRecipe：从网页图片提取提示词，再用自己的照片生成同款

最近做了一个 Chrome 插件，叫 **LookRecipe （照样拍同款）**，已经上架 Chrome 商店，来 V 站分享一下。 它想解决一个具体的问题：看到一张喜欢的照片，想知道它的光线、构图和配色是怎么回事，再把这些特点用到自己的创作里。 我把网页选图、图片分析、编辑提示词和 AI 生图串成了一个流程。 安装后，在普通网页上点击插件图标，选择图片或框选区域，就可以分析。结果会按主体、构图、镜头、光线、色调等维度拆成“摄影配方”，可以继续修改，也可以直接复制提示词。 拿到配方后，可以在插件里继续生成同款风格的图片。比如看中一张红西装人像里的硬光和蓝色背景，就可以围绕这些特点调整提示词，再加入自己的人物参考。 目前还做了几个配套功能： - **人物卡**：上传 1–5 张同一个人的照片，创建可复用的人物参考。 - **四宫格套图**：生成一张 2×2 摄影套图，再对喜欢的单张做高清增强。 - **本地素材库**：收集参考图，整理分析记录和生成结果，方便之后继续改。 - **中英文界面**：可以在设置中切换，不用自己配置 API Key 。 收费也提前说清楚： 本地收集免费，不需要登录。AI 分析和生图需要 Google 登录，首次登录赠送 3 点，够体验 3 次图片分析；单张生图需要 10 点。 其他消耗是：人物卡 20 点，四宫格套图 20 点，高清增强每张 10 点。失败或取消的云端任务不扣点。 点数包目前是 **4.99 美元 / 120 点、9.99 美元 / 300 点、19.99 美元 / 700 点**，一次性购买，购买的点数永久有效，没有自动续费。 数据处理方面，进行 AI 分析或生图时，提交的图片和提示词会发送到云端及 OpenAI 处理。素材和历史在本机管理，账号、点数和订单可以跨设备使用。 分析出的光线、焦段等属于 AI 根据画面做的推测，生成效果也会随参考图和提示词变化。如果你平时会收集人像、穿搭或摄影风格参考，可以看看这个流程是否用得上。 - [官网]( ) - [Chrome 商店安装]( ) - [59 秒中文演示]( ) 欢迎试用，有选图不顺手、分析不够具体，或者生图效果不理想的地方，可以直接在楼里反馈。

https://www.v2ex.com/t/1255093

## 做了个抖音运营助手 Chrome 扩展：推荐流挂机 + 关键词搜索 + AI 神评

开发者自荐：我是波波抖音内容运营助手的开发者，下面讲怎么做、怎么用，不讲流量保证。 ### 1. 痛点：抖音日常运营，重复动作太多 做抖音号，每天都要刷推荐流、找对标视频、回评论、盯意向客户。动作本身不难，难在每天重复、耗时，而且人工一累就容易漏掉真正想买的人。 ### 2. 它是怎么做到的 - 跑在 Chrome / Edge 的抖音网页版里，不用输入抖音账号密码，也不用导 Cookie ，在你已登录的官方页面里执行。 - 推荐流挂机：自动浏览推荐流，模拟完播停留和点赞、收藏，停留时长和互动概率可以自己调。 - 关键词搜索获客：多关键词队列搜索，结合抖音官方排序和视频时长筛选，批量定位对标视频，自动记录已处理视频，不重复处理。 - 评论区截流：在视频评论区直接发表评论，或在同行视频的高意向评论下做楼中楼回复，带包含词 / 排除词过滤。 - 云端大模型 AI 神评：结合视频标题和前排热评生成评论，支持自定义 Prompt 和语气，也支持纯文本话术库。 - 拟人节奏：随机停留、随机互动概率，遇到验证码或频控提示会平稳暂停，不硬冲。 - 记录都在本地，可以导出 CSV 复盘。 ### 3. 四种预设模式 - 稳健养号：新号、提权防风控 - 智能获客：均衡高效，转化首选 - 爆客冲量：老号抢量 - 极速测试：快速验证策略 模式可以一键切换，里面的停留、互动、AI 风格也可以自己再调。 ### 4. 安装与价格 - 安装：官网下载离线安装包 ZIP ，解压后在浏览器扩展管理里开开发者模式，加载已解压的扩展程序。 - 试用版免费：单次最多处理 15 个视频、单次挂机限时 15 分钟、每日 20 次免费 AI 神评。 - 永久专业版：前 50 名早鸟 ¥19.9 终身买断，名额满后恢复 ¥19.9/月。支付宝直付，付款成功后页面显示激活码，一机一码，粘贴到扩展里激活。剩余名额官网没给实时数字，草稿不写还剩多少名。 ### 5. 限制，一定要先说清楚 - 不能保证零风险，也不能绕过平台审核和平台规则。 - 请遵守抖音社区公约，合理设置单日互动上限，不滥发引流广告，不在短时间内超大频次操作。 - 自动化工具仍有账号风险，节奏和数量要按自己账号的实际情况调，不要无脑拉满。 ### 6.

https://www.v2ex.com/t/1255092

## 做了个 AI 视频小工具：一张照片生成「无边泳池火烈鸟」视频

大家好，做了一个叫 Infinity Pool AI 的小工具，来 V 站分享一下。 它做的事情很具体：上传一张单人照片，生成一段人物躺在火烈鸟泳圈上、低头玩手机、慢慢漂向无边泳池边缘，最后被一只手推回来的视频。也就是那个 “Hey bro / Thank you bro” 的泳池梗。 网站：Infinity Pool AI 首页放了参考照片和实际生成的成片，也可以直接看这个 12 秒视频示例。 目前专注这一个固定场景，视频长度是 12 秒，支持 480p 和 720p ，输出竖屏 MP4 。使用时选照片和清晰度就可以，不用自己写提示词。 技术上用的是 TanStack Start + React ，部署在 Cloudflare Workers ，用户、订单和积分数据存 D1 ，支付接 Stripe 。底层调用视频生成服务，站点这边主要负责场景模板、照片上传、异步任务管理和积分结算。 这一版也整理了登录、购买和生成之间的流程。选好照片再登录，照片和清晰度会保留；从套餐入口进去，也会保留原来选的套餐。生成前先确认积分费用，完成的视频可以在“我的视频”里下载，右上角能查看剩余积分。任务生成失败时，会退回对应的生成积分。 收费方式是一次性购买积分包，没有订阅，购买的积分不过期。套餐价格是： - $9.90：100 积分，可生成 1 条 480p 视频。 - $19.90：300 积分，可生成 3 条 480p ，或 1 条 720p 视频。 - $29.90：500 积分，可生成 5 条 480p ，或 2 条 720p 视频。 480p 每次使用 100 积分，720p 每次使用 250 积分，两种清晰度共用余额，通过 Stripe 用美元结算。 AI 生成的人物相似度和动作细节会有波动，首页可以直接看实际效果，再决定是否购买。 想请 V 友帮忙看看：这种固定场景的视频工具，大家更看重人物相似度、生成速度，还是可选场景？页面上如果有没说明白、让人不知道下一步怎么操作的地方，也欢迎指出来。

https://www.v2ex.com/t/1255089

## 倒行逆施 - 纯文字校友社区 - 时光校友圈

受 V2EX 启发，短视频时代，不见得文字内容没有机会。 时光校友圈，欢迎大家体验！

https://www.v2ex.com/t/1255079