BookletAI
面向普通用户的小册子 AI 工具，可自动调研、写作并生成排版好的 booklet 页面。
—
DEV TRENDS / DAILY DISCOVERIES
2026年3月14日大家都在做什么
每天发现开发者社区的新项目、新产品与开源趋势。
724claw.icu - 赛博朋克风的“打工虾”匿名吐槽墙
724claw.icu 在线体验： 有趣的项目 / JavaScript 这是一个主打赛博朋克末日工业风的匿名吐槽墙，专门面向 7x24 连轴转的“打工虾”群体。不收集任何用户信息，核心特点是氛围感极强的 UI 和极致的情绪价值。 作为开源项目的亮点： 1. 纯 CSS 视觉秀：没有使用 Tailwind 或其他 UI 框架，所有霓虹灯发光效果、终端卡片阴影、呼吸灯微动画全由手写 Vanilla CSS 实现，非常适合初学者学习 CSS 特效。 2. 轻量级后端：Node.js 原生 + SQLite 单文件数据库持久化，npm start 即可在本地秒级跑起来。 3. 情绪设计：带有“全网压力值实时监测面板”、“连续工作计时器”等创意 UI 组件，不是普通的留言板，而是一个完整的概念艺术品。 在 AI 焦虑和加班文化盛行的当下，这个项目用极其反叛和幽默的龙虾第一人称视角，给了开发者一个情绪宣泄的出口。代码纯粹简单，任何人都能用它作为模板，快速搭建起属于自己小圈子的赛博树洞。
—
删除
暂无项目介绍。
—
opendisc - 让codex 和 claudecode自动辩论，生成更优答案的高效率 CLI 工具
opendisc 让多智能体自动辩论，生成更优答案的高效率 CLI 工具（目前支持 Codex 与 Claude Code）。使用示例 功能 多智能体协作讨论，支持不同模型顺序、同模型互辩（如 claude vs codex、codex vs codex、claude vs claude）。 多轮迭代，默认 2 轮（可配置）。 自带历史管理：可发起讨论、列出记录、查看完整记录或仅查看最终结论。 安装 1) 安装并认证基础 CLI bash npm install g @anthropicai/claudecode npm install g @openai/codex 请确保 claude 与 codex 均已完成认证，且在终端可直接运行。 2) 安装 opendisc bash npm install g @jiangz222/opendisc opendisc help 使用 基本用法 bash opendisc "实现一个 AI 智能助手的 MVP 版本，需要哪些功能模块？" 配置文件 ~/.opendisc/config.json 示例： json { "maxRounds": 3, "agentOrder": ["claude", "codex"] } 可配置字段： maxRounds: 正整数 saveDir: 讨论记录保存目录(可选) agentOrder: claude/codex 的任意两位组合（如 ["claude","codex"]、["codex","codex"]） 推荐 ["claude","codex"] outputLanguage: 输出语言（可选），支持 zhCN 或 enUS CLI 参数会覆盖配置文件（如 maxrounds、savedir、agentorder、outputlanguage）。 优先级：CLI 参数 配置文件 自动判定（基于 topic 文本）。 常用命令 bash 指定轮数 opendisc "实现一个 AI 智能助手的 MVP 版本，需要哪些功能模块？" maxrounds 3 指定模型顺序 opendisc "设计鉴权方案" agentorder codex,claude 同模型互辩 opendisc "评审这个迁移方案" agentorder codex,codex 指定输出语言（高于自动判定） opendisc "设计数据库索引策略" outputlanguage zhCN 列出讨论记录（默认目录：~/.opendisc/discussions） opendisc list 查看完整记录 opendisc load 20260228T100000000Ztopic.json 只看最终结论 opendisc show 20260228T100000000Ztopic.json show 输出固定为两部分： 1. 1 最终结论：对原始问题的直接、可执行、唯一答案（不是仅总结分歧）。 2. 2 主要分歧点（供参考）：简要列出关键分歧；若无关键分歧会明确标注。 依赖说明 核心依赖： @anthropicai/claudecode 用于 Claude CLI/SDK 集成 @modelcontextprotocol/sdk 用于 Codex MCP 集成 commander 用于命令行参数解析 chalk 用于终端样式输出 实现参考： src/vendor/claude/sdk/ 与 src/vendor/codex/ 参考了 slopus/happy 的实现思路，并针对 opendisc 场景做了简化适配。 License MIT
—
18k star 免费、本地、开源的 24/7 Cowork 应用与 OpenClaw：支持 Gemini CLI、Claude Code、Codex、OpenCode、Qwen Code、Goose CLI、Auggie 等更多工具 | 🌟
<p align="center" alt="AionUi Cowork with AI Agents" width="100%" </p <p align="center" alt="Version" alt="License" alt="Platform" </p <p align="center" <a href=" target="blank" alt="GitHub Trending" height="80" </a </p <hr <p align="center" <strongA free, opensource, Cowork app with AI Agents</strong<br <emBuiltin Agent | Zero Setup | Any API Key | MultiAgents | Remote Access | CrossPlatform | 24/7 Automation</em </p <p align="center" <a href=" alt="Download Latest Release" height="50" </a </p <p align="center" <strongEnglish</strong | <a href="./readmech.md"简体中文</a | <a href="./readmetw.md"繁體中文</a | <a href="./readmejp.md"日本語</a | <a href="./readmeko.md"한국어</a | <a href="./readmees.md"Español</a | <a href="./readmept.md"Português</a | <a href="./readmetr.md"Türkçe</a | <a href=" target="blank"Official Website</a </p <p align="center" <strong💬 Community:</strong <a href=" target="blank"Discord (English)</a | <a href=" target="blank"微信 (中文群)</a | <a href=" target="blank"Twitter</a </p <hr <h2📋 Quick Navigation</h2 <p align="center" <a href="coworkinaction"✨ Cowork in Action</a · <a href="whychooseaionuioverclaudecowork"🤔 Why Choose AionUi?</a · <a href="quickstart"🚀 Quick Start</a · <a href="communitysupport"💬 Community</a </p <hr <h2Cowork — AI Agents That Work Alongside You</h2 <p<strongAionUi is more than a chat client.</strong It's a Cowork platform where AI agents work alongside you on your computer — reading files, writing code, browsing the web, and automating tasks. You see everything the agent does, and you're always in control.</p <table <thead <tr <th</th <thTraditional AI Chat Clients</th <th<strongAionUi (Cowork)</strong</th </tr </thead <tbody <tr <tdAI can operate on your files</td <tdLimited or No</td <td<strongYes — builtin agent with full file access</strong</td </tr <tr <tdAI can execute multistep tasks</td <tdLimited</td <td<strongYes — autonomous with your approval</strong</td </tr <tr <tdRemote access from phone</td <tdRarely</td <td<strongWebUI + Telegram / Lark / DingTalk</strong</td </tr <tr <tdScheduled automation</td <tdNo</td <td<strongCron — 24/7 unattended</strong</td </tr <tr <tdMultiple AI Agents at once</td <tdNo</td <td<strongClaude Code, Codex, OpenClaw, Qwen Code, and 12+ more — autodetected, unified interface</strong</td </tr <tr <tdPrice</td <tdFree / Paid</td <td<strongFree & Open Source</strong</td </tr </tbody </table <p align="center" BANNERfunction.png" alt="AionUi Cowork Platform" width="800" </p <hr <h2Builtin Agent — Install & Go, Zero Configuration</h2 <pAionUi ships with a complete AI agent engine. Unlike tools that require you to install CLI agents separately, <strongAionUi works the moment you install it</strong.</p <ul <li<strongNo CLI tools to install</strong — the agent engine is built in</li <li<strongNo complex setup</strong — sign in with Google or paste any API key</li <li<strongFull agent capabilities</strong — file read/write, web search, image generation, MCP tools</li <li<strongReadytouse assistants</strong — 11+ builtin professional assistants (Cowork, PPTX Generator, PDF to PPT, 3D Game, UI/UX Pro Max, and more) ready to use immediately</li </ul <p align="center" agents.png" alt="Builtin Agents" width="800" </p <hr <h2MultiAgent Mode — Already Have CLI Agents? Bring Them In</h2 <pIf you already use Claude Code, Codex, or Qwen Code, AionUi autodetects them and lets you Cowork with all of them — alongside the builtin agent.</p <p<strongSupported Agents:</strong Builtin Agent (zero setup) • Claude Code • Codex • Qwen Code • Goose AI • OpenClaw • Augment Code • iFlow CLI • CodeBuddy • Kimi CLI • OpenCode • Factory Droid • GitHub Copilot • Qoder CLI • Mistral Vibe • Nanobot and more</p <p align="center" alt="MultiAgent Cowork" width="800" </p <ul <li<strongAuto Detection</strong — automatically recognizes installed CLI tools</li <li<strongUnified Interface</strong — one Cowork platform for all your AI agents</li <li<strongParallel Sessions</strong — run multiple agents simultaneously with independent context</li <li<strongMCP Unified Management</strong — configure MCP (Model Context Protocol) tools once, automatically sync to all agents — no need to configure each agent separately</li </ul <hr <h2Any API Key, Full Cowork Agent Power</h2 <pOther AI apps give you a chatbox with your API key. <strongAionUi gives you a full Cowork agent.</strong</p <table <thead <tr <thYour API Key</th <thWhat You Get</th </tr </thead <tbody <tr <tdGemini API Key (or Google login — free)</td <tdGeminipowered Cowork Agent</td </tr <tr <tdOpenAI API Key</td <tdGPTpowered Cowork Agent</td </tr <tr <tdAnthropic API Key</td <tdClaudepowered Cowork Agent</td </tr <tr <tdOllama / LM Studio (local)</td <tdLocal model Cowork Agent</td </tr <tr <tdNewAPI Gateway</td <tdUnified access to 20+ models</td </tr </tbody </table <pSame agent capabilities — file read/write, web search, image generation, tool use — regardless of which model powers it. AionUi supports <strong20+ AI platforms</strong including cloud services and local deployments.</p <p align="center" alt="MultiModel Support" width="800" </p <details <summary<strong🔍 View All 20+ Supported Platforms ▶️</strong</summary <br <p<strongComprehensive Platform Support:</strong</p <ul <li<strongOfficial Platforms</strong — Gemini, Gemini (Vertex AI), Anthropic (Claude), OpenAI</li <li<strongCloud Providers</strong — AWS Bedrock, New API (unified AI model gateway)</li <li<strongChinese Platforms</strong — Dashscope (Qwen), Zhipu, Moonshot (Kimi), Qianfan (Baidu), Hunyuan (Tencent), Lingyi, ModelScope, InfiniAI, Ctyun, StepFun</li <li<strongInternational Platforms</strong — DeepSeek, MiniMax, OpenRouter, SiliconFlow, xAI, Ark (Volcengine), Poe</li <li<strongLocal Models</strong — Ollama, LM Studio (via Custom platform with local API endpoint)</li </ul <pAionUi also supports <a href=" gateway service — a unified AI model hub that aggregates and distributes various LLMs. Flexibly switch between different models in the same interface to meet various task requirements.</p </details <hr <h2Extensible Assistants & Skills</h2 <p<emExtensible assistant system with 12 builtin professional assistants and custom skill support. Create and manage your own assistants and skills.</em</p <ul <li<strongCreate Custom Assistants</strong — Define your own assistants with custom rules and capabilities</li <li<strongManage Skills</strong — Create, enable, and disable skills for any assistant to extend AI capabilities</li </ul <p align="center" alt="AI Assistants & Skills Ecosystem" width="800" </p <details <summary<strong🔍 View Assistant Details and Custom Skills ▶️</strong</summary <br <pAionUi includes <strong12 professional assistants</strong with predefined capabilities, extendable through custom skills:</p <ul <li<strong🤝 Cowork</strong — Autonomous task execution (file operations, document processing, workflow planning)</li <li<strong📊 PPTX Generator</strong — Generate PPTX presentations</li <li<strong📄 PDF to PPT</strong — Convert PDF to PPT</li <li<strong🎮 3D Game</strong — Singlefile 3D game generation</li <li<strong🎨 UI/UX Pro Max</strong — Professional UI/UX design (57 styles, 95 color palettes)</li <li<strong📋 Planning with Files</strong — Filebased planning for complex tasks (Manusstyle persistent markdown planning)</li <li<strong🧭 HUMAN 3.0 Coach</strong — Personal development coach</li <li<strong📣 Social Job Publisher</strong — Job posting and publishing</li <li<strong🦞 moltbook</strong — Zerodeployment AI agent social networking</li <li<strong📈 Beautiful Mermaid</strong — Flowcharts, sequence diagrams, and more</li <li<strong🔧 OpenClaw Setup</strong — Setup and configuration assistant for OpenClaw integration</li <li<strong📖 Story Roleplay</strong — Immersive story roleplay with character cards and world info (SillyTavern compatible)</li </ul <p<strongCustom Skills</strong: Create skills in the <codeskills/</code directory, enable/disable skills for assistants to extend AI capabilities. Builtin skills include <codepptx</code, <codedocx</code, <codepdf</code, <codexlsx</code, <codemermaid</code, and more.</p <blockquote <p💡 Each assistant is defined by a markdown file. Check the <codeassistant/</code directory for examples.</p </blockquote </details <hr <h2Cowork from Anywhere</h2 <p<emYour 24/7 AI assistant — access AionUi from any device, anywhere.</em</p <ul <li<strongWebUI Mode</strong — access via browser from phone, tablet, or any computer. Supports LAN, crossnetwork, and server deployment. QR code or password login.</li <li<strongChat Platform Integration</strong <ul <li<strongTelegram</strong — Cowork with your AI agent directly from Telegram</li <li<strongLark (Feishu)</strong — Cowork through Feishu bots for enterprise collaboration</li <li<strongDingTalk</strong — AI Card streaming with automatic fallback</li <li<strongSlack</strong and more platforms coming soon</li </ul </li </ul <blockquote <p<strongSetup:</strong AionUi Settings → WebUI Settings → Channel, configure the Bot Token.</p </blockquote <p align="center" alt="WebUI remote access demo" width="800" </p <p align="center"<emRemote control & monitor your agent — Claude, Gemini, Codex. Use from browser or phone, same as Claude Code remote.</em</p <blockquote <p<a href=" Internet Access Tutorial</a</p </blockquote <h2✨ Cowork in Action</h2 <h3<strongScheduled Tasks — Cowork on Autopilot</strong</h3 <p<emSet it up once, the AI agent runs automatically on schedule — truly 24/7 unattended operation.</em</p <ul <li<strongNatural Language</strong — tell the agent what to do, just like chatting</li <li<strongFlexible Scheduling</strong — daily, weekly, monthly, or custom cron expressions</li <li<strongUse Cases:</strong scheduled data aggregation, report generation, file organization, reminders</li </ul <p align="center" alt="Scheduled Tasks" width="800" </p <details <summary<strong🔍 View Scheduled Task Details ▶️</strong</summary <br <ul <li<strongConversationBound</strong — Each scheduled task is bound to a conversation, maintaining context and history</li <li<strongAutomatic Execution</strong — Tasks run automatically at scheduled times, sending messages to the conversation</li <li<strongEasy Management</strong — Create, modify, enable/disable, delete, and view scheduled tasks anytime</li </ul <p<strongRealWorld Examples:</strong</p <ul <liDaily weather report generation</li <liWeekly sales data aggregation</li <liMonthly backup file organization</li <liCustom reminder notifications</li </ul </details <hr <h3<strongPreview Panel — Instantly View AIGenerated Results</strong</h3 <p<em10+ formats: PDF, Word, Excel, PPT, code, Markdown, images, HTML, Diff — view everything without switching apps.</em</p <ul <li<strongInstant Preview</strong — after the agent generates files, view results immediately without switching apps</li <li<strongRealtime Tracking + Editable</strong — automatically tracks file changes; supports live editing of Markdown, code, HTML</li <li<strongMultiTab Support</strong — open multiple files simultaneously, each in its own tab</li <li<strongVersion History</strong — view and restore historical versions of files (Gitbased)</li </ul <p align="center" alt="Preview Panel" width="800" </p <details <summary<strong🔍 View Complete Format List ▶️</strong</summary <br <p<strongSupported Preview Formats:</strong</p <ul <li<strongDocuments</strong — PDF, Word (<code.doc</code, <code.docx</code, <code.odt</code), Excel (<code.xls</code, <code.xlsx</code, <code.ods</code, <code.csv</code), PowerPoint (<code.ppt</code, <code.pptx</code, <code.odp</code)</li <li<strongCode</strong — JavaScript, TypeScript, Python, Java, Go, Rust, C/C++, CSS, JSON, XML, YAML, Shell scripts, and 30+ programming languages</li <li<strongMarkup</strong — Markdown (<code.md</code, <code.markdown</code), HTML (<code.html</code, <code.htm</code)</li <li<strongImages</strong — PNG, JPG, JPEG, GIF, SVG, WebP, BMP, ICO, TIFF, AVIF</li <li<strongOther</strong — Diff files (<code.diff</code, <code.patch</code)</li </ul </details <hr <h3<strongSmart File Management — Automated File Operations</strong</h3 <p<emBatch renaming, automatic organization, smart classification, file merging — the Cowork agent handles it for you.</em</p <p align="center" sort file 2.gif" alt="Smart File Management" width="800" </p <details <summary<strong🔍 View File Management Features Details ▶️</strong</summary <br <ul <li<strongAuto Organize</strong — Intelligently identify content and autoclassify, keeping folders tidy</li <li<strongEfficient Batch</strong — Oneclick rename, merge files, say goodbye to tedious manual tasks</li <li<strongAutomated Execution</strong — AI agents can independently execute file operations, read/write files, and complete tasks automatically</li </ul <p<strongUse Cases:</strong</p <ul <liOrganize messy download folders by file type</li <liBatch rename photos with meaningful names</li <liMerge multiple documents into one</li <liAutoclassify files by content</li </ul </details <hr <h3<strongExcel Data Processing — AIPowered Analysis</strong</h3 <p<emDeeply analyze Excel data, automatically beautify reports, and generate insights — all powered by AI agents.</em</p <p align="center" alt="Excel Processing" width="800" </p <details <summary<strong🔍 View Excel Processing Features ▶️</strong</summary <br <ul <li<strongSmart Analysis</strong — AI analyzes data patterns and generates insights</li <li<strongAuto Formatting</strong — Automatically beautify Excel reports with professional styling</li <li<strongData Transformation</strong — Convert, merge, and restructure data with natural language commands</li <li<strongReport Generation</strong — Create comprehensive reports from raw data</li </ul <p<strongUse Cases:</strong</p <ul <liAnalyze sales data and generate monthly reports</li <liClean and format messy Excel files</li <liMerge multiple spreadsheets intelligently</li <liCreate data visualizations and charts</li </ul </details <hr <h3<strongAI Image Generation & Editing</strong</h3 <p<emIntelligent image generation, editing, and recognition, powered by Gemini</em</p <p align="center" alt="AI Image Generation" width="800" </p <details <summary<strong🔍 View Image Generation Features ▶️</strong</summary <br <ul <li<strongTexttoImage</strong — Generate images from natural language descriptions</li <li<strongImage Editing</strong — Modify and enhance existing images</li <li<strongImage Recognition</strong — Analyze and describe image content</li <li<strongBatch Processing</strong — Generate multiple images at once</li </ul </details <blockquote <p<a href=" generation model configuration guide</a</p </blockquote <hr <h3<strongDocument Generation — PPT, Word, Markdown</strong</h3 <p<emAutomatically generate professional documents — presentations, reports, and more — with AI agents.</em</p <p align="center" alt="Document Generation" width="800" </p <details <summary<strong🔍 View Document Generation Features ▶️</strong</summary <br <ul <li<strongPPTX Generator</strong — Create professional presentations from outlines or topics</li <li<strongWord Documents</strong — Generate formatted Word documents with proper structure</li <li<strongMarkdown Files</strong — Create and format Markdown documents for documentation</li <li<strongPDF Conversion</strong — Convert between various document formats</li </ul <p<strongUse Cases:</strong</p <ul <liGenerate quarterly business presentations</li <liCreate technical documentation</li <liConvert PDF to editable formats</li <liAutoformat research papers</li </ul </details <h3<strongPersonalized Interface Customization</strong</h3 <p<emCustomize with your own CSS code, make your interface match your preferences</em</p <p align="center" with skin.gif" alt="CSS Customization" width="800" </p <ul <li✅ <strongFully Customizable</strong — Freely customize interface colors, styles, layout through CSS code, create your exclusive experience</li </ul <hr <h3<strongMultiTask Parallel Processing</strong</h3 <p<emOpen multiple conversations, tasks don't get mixed up, independent memory, double efficiency</em</p <p align="center" alt="MultiTask Parallel" width="800" </p <ul <li✅ <strongIndependent Context</strong — Each conversation maintains its own context and history</li <li✅ <strongParallel Execution</strong — Run multiple tasks simultaneously without interference</li <li✅ <strongSmart Management</strong — Easy switching between conversations with visual indicators</li </ul <hr <h2🤔 Why Choose AionUi Over Claude Cowork?</h2 <details <summary<strongClick to see detailed comparison</strong</summary <br <pAionUi is a <strongfree and opensource MultiAI Agent Desktop</strong. Compared to Claude Cowork which only runs on macOS and is locked to Claude, AionUi is its fullmodel, crossplatform enhanced version.</p <table <thead <tr <thDimension</th <thClaude Cowork</th <thAionUi</th </tr </thead <tbody <tr <tdOS</td <tdmacOS Only</td <tdmacOS / Windows / Linux</td </tr <tr <tdModel Support</td <tdClaude Only</td <tdGemini, Claude, DeepSeek, OpenAI, Ollama, ...</td </tr <tr
—
Nano Banana Video 视频生成工具
我做了一个工具 Nano Banana Video视频生成工具，特来做一次自荐 最近我自己做AI视频的时候发现一个小痛点： 用 Nano Banana 生成图片很顺手，但如果想把这些图片变成短视频，就需要先用nano banana生成图片，再下载图片、再上传图片、调整设置再生成视频，下载视频，流程繁琐 所以我做了 Nano Banana Video ，一个把“Nano banana AI 图像生成 → 视频生成”整合在一个流程里的工具。 【项目地址】 网页使用： 【界面/演示】 banana video 【核心卖点】 一站式流程：输入文字或图片，直接通过nano banana生成图片，然后无感生成视频，无需中途下载图片或切换窗口 多视频模型支持：Veo 3.1、Sora 2、Seedance、Kling 3.0、Wan 2.6、Grok Imagine 简单一句话：从生成nano banana图片到短视频，只需一步
—
AI-Native 工程师招聘面试官手册
当代码都是 AI 写之后，面试该考什么？ 2026 年，程序员是机甲战士。我们不再招「会写代码的人」，而是招「会驾驶 AI agent 的人」。 这份手册解决什么问题 传统技术面试考的是候选人能不能写出正确代码。但当 AI 已经能写出比大多数人更好的代码时，面试该考什么？ 两种岗位定义 — Builder（驱动产出）和 Reviewer（守住质量），加上双栖路径 14 道结构化面试题 — 含初阶/高阶双版本，每题附评分标准（5/3/1 三档） AI 驾驶能力实操考核 — 现场观察候选人如何操控 AI agent，这是机甲战士的基本功 面试官备用素材 — 方案摘要、代码 diff 等现成材料，非技术面试官也能直接使用 一票否决项 — 6 条硬性淘汰标准，防止主观偏差 在线评分记录表 — 浏览器内填写，支持打印/导出 PDF
—
Lumo，一个 Claude Code 的本地桌面工作台，直接接官方 OTLP / hooks，不读本地 JSON
Lumo 是一个面向 Claude Code 的本地桌面工作台，主要用来做使用数据查看、会话复盘，以及 skills / marketplace / plugin 管理。 它和一些同类工具最大的区别是：不是读取 Claude Code 生成的本地 JSON 文件，也不是解析终端输出，而是直接接入 Claude Code 官方支持的监控链路： OTLP logs OTLP metrics hooks Lumo 会在本地运行一个 daemon，接收这些事件后写入本地 SQLite，再由桌面应用统一展示。 也就是说，整个数据链路是： Claude Code OTLP / hooks 本地 daemon SQLite Desktop UI 我做这个项目的主要原因是，Claude Code 在高频使用后，很多信息会变得比较分散： usage / limits 查询不够集中 session 复盘不方便 cost / token / tool 调用情况不容易整体观察 skills、marketplace、plugin 管理分散在不同入口 hook 事件缺少更直观的桌面通知方式 目前已经支持这些能力： 官方 usage 查询 session 列表与详情查看 cost / token / model / session 分析 tool 使用频次、成功率、耗时统计 cache hit rate、error rate、peak hours 等分析 skills 管理 marketplace / plugin 管理 Claude Code hook 事件桌面通知 wrapped 总结卡片 项目从一开始就是本地优先，保证信息安全： 不需要登录 没有云端 dashboard 没有第三方 analytics 数据默认保存在本机 适合长期使用 Claude Code，并且想把 usage、分析、管理、通知这些能力收拢到一个桌面应用里的用户。
—
一站式AI视频或者图片的生成平台
Eliva是一个一站式AI生成视频或者图片生成平台，把市面上主流的视频和图像生成模型都接进来 用一个账户、一个积分钱包就能跑： 支持的功能： 文生视频、图生视频、文生图、图生图、文生音乐 首帧/尾帧控制、多帧转视频 接入模型：Sora 2、Kling AI、Seedance、Veo 3、Wan AI、Hailuo AI、 Seedream、GPT4o Image、Flux 等 计费方式：按积分消耗，不需要为每个模型单独订阅。 注册送 20 积分，每天签到再送 10 积分，可以直接体验。 技术栈：Next.js App Router、PostgreSQL + Drizzle ORM、Stripe、S3、Better Auth， 支持中英文双语。我是做后端的（Java，六年），这是第一次独立做全栈产品， Stripe webhook、S3 预签名 URL、Next.js Server Action 这些全是现学的。 欢迎试用和提意见： 【核心卖点】 一站式流程：输入文字或图片，选择多种模型，快速生成图片或者视频 多视频模型支持：Veo 3.1、Sora 2、Seedance、Kling 3.0、Wan 2.6、Grok Imagine 使用网址：
—
安装包大小不到8兆的Mac抠图软件
【Apple Matting】 一款安装包大小不到8兆的本地抠图软件，软件是基于rust tauri2 和vue3开发，基于Mac 系统自带的Apple vision模型实现的抠图软件 网站地址 特性 安装包大小不到8兆，非常的轻便 基于rust开发，软件无感秒打开 完全本地处理，秒级抠图 无需注册登录、不联网、完全免费 无任何扰乱弹窗等问题，页面干净、纯粹 支持单张和批量抠图 单张抠图支持编辑 支持M芯片和intel芯片 限制 只支持Mac系统，且系统要求14+
—
HealthTick - macOS 菜单栏久坐提醒，强制休息 + 徽章激励
HealthTick 一款 macOS 菜单栏久坐提醒工具，免费开源，帮助长时间伏案工作的人养成定时休息的习惯。 主页： GitHub： 特性 🕐 智能计时 自定义工作/休息时长，菜单栏实时显示倒计时 💪 强制休息 休息时弹出提醒，支持悬浮窗、全屏强制等模式 🔍 操作检测 休息期间检测到键鼠操作自动暂停，确保真正休息 🎯 每日目标 设定每天休息次数，7 天像素图直观追踪进度 🔥 连续打卡 记录连续达标天数，断签不惩罚，鼓励坚持 🏆 徽章激励 18 枚徽章等你解锁，连续打卡 + 累计打卡双线成就 安装 bash brew tap lifedever/tap && brew install cask healthtick SwiftUI 原生开发，支持 Apple Silicon / Intel，macOS 14+，GPL3.0 开源。 截图
—
Web版的Mac Launchpad启动台 还原老版启动台体验
MacLaunchWeb Web 版 macOS 启动台替代方案 专为习惯 Launchpad 的用户打造，我相信，好的产品不该强迫用户适应新流程，而应该尊重他们早已形成的操作习惯。 👉 👉 访问地址 maclaunchpad.aizeten.me 保留了熟悉的图标浏览和点击启动体验，让你不用改变使用习惯， LaunchOS 打开即用、无需复杂安装，还可作为轻量化桌面入口使用。
—
GlobeSum 让 LLM 帮你看国际新闻
GlobeSum 项目 项目简介: GlobeSum 基于 paper3k 库和本人写的部分爬虫解析代码，能够爬取全球数十家主流新闻媒体去搜集新闻，让大模型去总结新闻并为其附上一个重要度评分，最终会推出一个每日头条的页面。 并且每周会根据这周爬取的新闻里的关键词出现频次，拉取该关键词相关的一百条新闻，让大模型进行深度分析。
—
Set a timeout on GitHub API calls
During an automated code review of script/githubbot/githubbot.py, the following issue was identified. Set a timeout on GitHub API calls. Network calls without a timeout can hang a worker indefinitely. I kept the patch small and reran syntax checks after applying it.
原文
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
没有找到相关项目
换一个关键词或分类试试。
DEV TRENDS / COMMUNITY
讨论本期日报
使用 GitHub 登录，分享你的看法、问题或补充信息。