[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fD95647zpMPspJL7xZQpdudW9GEawHmHGgePLb3chVmU":3,"tool-scrapling":16},{"siteName":4,"siteDescription":5,"siteUrl":6,"contactEmail":7,"socialLinks":8,"analyticsCode":10,"seo":11},"AllAI App","Call 500+ AI models through one compliant API KEY. Discover 35,000+ AI tools across 21 categories. Compare, test all ai app in one platform.","https:\u002F\u002Fallaiapp.com","contact@allaiapp.com",{"twitter":9,"github":9,"discord":9,"linkedin":9,"youtube":9},"","\u003C!-- Google tag (gtag.js) -->\n\u003Cscript async src=\"https:\u002F\u002Fwww.googletagmanager.com\u002Fgtag\u002Fjs?id=G-3NJE4SY4LE\">\u003C\u002Fscript>\n\u003Cscript>\n  window.dataLayer = window.dataLayer || [];\n  function gtag(){dataLayer.push(arguments);}\n  gtag('js', new Date());\n\n  gtag('config', 'G-3NJE4SY4LE');\n\u003C\u002Fscript>\n\u003C!-- Baidu Tongji (hm.js) -->\n\u003Cscript>\nvar _hmt = _hmt || [];\n(function() {\n  var hm = document.createElement(\"script\");\n  hm.src = \"https:\u002F\u002Fhm.baidu.com\u002Fhm.js?ee51358c8d4491dea1125cfc1d47a169\";\n  var s = document.getElementsByTagName(\"script\")[0];\n  s.parentNode.insertBefore(hm, s);\n})();\n\u003C\u002Fscript>",{"titleTemplate":12,"metaKeywords":13,"ogImage":14,"twitterSite":9,"defaultRobots":15,"googleVerify":9,"bingVerify":9,"baiduVerify":9,"organizationLogo":14},"%s | AllAI App","AI KEY,AI API KEY,Deepseek Key,Chatgpt Key,OpenAi Key,AI tools, AI directory, AI API gateway, AI models, artificial intelligence,","https:\u002F\u002Fallaiapp.com\u002Fog-image.png","index, follow",{"tool":17,"related":220},{"slug":18,"name":19,"cnName":20,"cn_name":20,"url":21,"pricingUrl":21,"github":21,"logo":9,"category":22,"subcategory":9,"categories":23,"pricing":24,"priceFrom":25,"accent":26,"platforms":27,"inputs":31,"outputs":34,"languages":38,"tags":41,"capabilities":50,"rating":25,"votes":25,"reviewCount":25,"saves":25,"monthlyVisits":25,"foundedYear":25,"founded":9,"developer":51,"status":52,"editorPick":53,"featured":53,"verified":53,"addedAt":54,"lastUpdated":54,"updatedAt":54,"en":55,"zh":137},"scrapling","Scrapling","Scrapling 自适应爬虫","https:\u002F\u002Fgithub.com\u002FD4Vinci\u002FScrapling","code",[22],"free",0,"#F59E0B",[28,29,30],"Python","CLI","Self-hosted",[32,33],"URL","Text",[35,36,37],"Data","Markdown","Code",[39,40],"中文","English",[42,43,44,45,46,47,48,49],"adaptive scraper","self-healing","anti-bot","cloudflare","playwright","MCP","自适应爬虫","反爬",[],"D4Vinci","active",false,"2026-09-13",{"tagline":56,"description":57,"features":58,"pros":83,"cons":89,"useCases":94,"pricingTiers":110,"faq":124},"The first adaptive, self-healing Python scraping library: it relocates elements after a site redesign and ships stealth fetchers that bypass Cloudflare-level anti-bot.","Scrapling solves the two costliest problems in production scraping: getting the response and keeping selectors alive. Its fast lxml-based parser supports CSS\u002FXPath\u002Fregex\u002Ftext queries; when a website changes layout, an adaptive mode (adaptive=True) uses element fingerprints — structure, text patterns, sibling relations — to automatically relocate the same element, so spiders are 'redesign-proof'. Three progressive fetcher tiers cover plain HTTP, full-browser Playwright and a StealthyFetcher (Camoufox-based) that hides CDP\u002FWebRTC leaks and clears Cloudflare Turnstile\u002FInterstitial automatically. It also offers a Scrapy-style spider framework with pause\u002Fresume and proxy rotation, sessions and an MCP server for AI extraction (BSD-3).",[59,62,65,68,71,74,77,80],{"title":60,"desc":61},"Self-healing selectors","adaptive=True re-finds elements after layout changes using fingerprints.",{"title":63,"desc":64},"Blazing parser","lxml-backed engine, dramatically faster than BeautifulSoup.",{"title":66,"desc":67},"Three fetcher tiers","Plain HTTP, full Playwright browser and stealth fetcher in one API.",{"title":69,"desc":70},"Anti-bot bypass","StealthyFetcher clears Cloudflare Turnstile and hides browser leaks.",{"title":72,"desc":73},"Spider framework","Scrapy-style crawlers with pause\u002Fresume and proxy rotation.",{"title":75,"desc":76},"Sessions","Persistent Fetcher\u002FStealthy sessions across requests.",{"title":78,"desc":79},"MCP server","Expose extraction to AI assistants via Model Context Protocol.",{"title":81,"desc":82},"Query flexibility","CSS, XPath, regex and text search over the same response.",[84,85,86,87,88],"Unique adaptive relocation saves constant selector maintenance.","Strong out-of-box anti-bot handling.","Fast parser with a clean progressive-enhancement API.","Sessions, spiders and proxies included.","Open source (BSD-3).",[90,91,92,93],"Stealth\u002Fbrowser tiers are heavier than plain requests.","Adaptive matching is best-effort on total rewrites.","Anti-bot arms race means occasional updates are needed.","Use only on sites and data you are permitted to collect.",[95,98,101,104,107],{"user":96,"desc":97},"Long-running scrapers","Keep collectors alive despite frequent site redesigns.",{"user":99,"desc":100},"Protected sites","Use StealthyFetcher where ordinary requests are blocked.",{"user":102,"desc":103},"Production pipelines","Combine fast HTTP with browser fallback in one codebase.",{"user":105,"desc":106},"AI extraction","Drive scraping through the MCP server from an assistant.",{"user":108,"desc":109},"Maintenance-light teams","Avoid rewriting selectors after every UI change.",[111,116,119],{"name":112,"price":113,"period":9,"features":114},"Open source (BSD-3)","Free",[115,30],"Full library",{"name":30,"price":113,"period":9,"features":117},[118],"Your infra\u002Fproxies",{"name":120,"price":121,"period":9,"features":122},"Optional integrations","Variable",[123],"Captcha\u002Fproxy providers if needed",[125,128,131,134],{"q":126,"a":127},"What makes Scrapling unique?","Its adaptive parser re-locates elements after a site changes, plus built-in stealth fetchers.",{"q":129,"a":130},"How does adaptive work?","It saves structural\u002Ftext fingerprints and similarity-matches them when selectors break.",{"q":132,"a":133},"Can it beat Cloudflare?","StealthyFetcher automates Turnstile\u002FInterstitial, though no tool bypasses everything forever.",{"q":135,"a":136},"Is it legal?","Only collect data you have the right to, respecting terms and law.",{"tagline":138,"description":139,"features":140,"pros":165,"cons":171,"useCases":176,"pricingTiers":192,"faq":207},"首个自适应、自愈的Python爬虫库：网站改版后能自动重新定位元素，内置隐秘抓取器可绕过 Cloudflare 级反爬。","Scrapling 解决生产爬虫最烧钱的两件事：拿到响应、让选择器一直有效。它基于 lxml 的高速解析器支持 CSS\u002FXPath\u002F正则\u002F文本查询；当网站改版时，自适应模式(adaptive=True)用元素指纹——结构、文本规律、兄弟节点关系——自动重新定位同一元素，让爬虫“对改版免疫”。三个渐进增强的抓取层覆盖纯HTTP、完整Playwright浏览器、以及基于 Camoufox 的 StealthyFetcher（隐藏 CDP\u002FWebRTC 泄漏、自动过 Cloudflare Turnstile\u002FInterstitial）。还提供类Scrapy蜘蛛框架（可暂停\u002F续跑、代理轮换）、会话管理，以及用于AI抽取的 MCP 服务（BSD-3）。",[141,144,147,150,153,156,159,162],{"title":142,"desc":143},"自愈选择器","adaptive=True靠指纹在改版后重新定位元素。",{"title":145,"desc":146},"极速解析","lxml引擎，远快于BeautifulSoup。",{"title":148,"desc":149},"三层抓取","纯HTTP\u002F完整Playwright\u002F隐秘浏览器同一API。",{"title":151,"desc":152},"反爬绕过","StealthyFetcher自动过Cloudflare、隐藏浏览器泄漏。",{"title":154,"desc":155},"蜘蛛框架","类Scrapy，可暂停续跑、轮换代理。",{"title":157,"desc":158},"会话管理","跨请求保持Fetcher\u002FStealthy会话。",{"title":160,"desc":161},"MCP服务","通过模型上下文协议让AI助手调用抽取。",{"title":163,"desc":164},"查询灵活","同一响应支持CSS\u002FXPath\u002F正则\u002F文本检索。",[166,167,168,169,170],"独有的自适应重定位，省掉大量选择器维护。","开箱即用的反爬能力强。","解析快、渐进增强API清晰。","会话\u002F蜘蛛\u002F代理齐备。","BSD-3开源。",[172,173,174,175],"隐秘\u002F浏览器层比纯请求重。","网站彻底重写时自适应只是尽力匹配。","反爬是军备竞赛，偶尔需更新。","只能采集获允许的站点与数据。",[177,180,183,186,189],{"user":178,"desc":179},"长期运行的爬虫","网站频繁改版也能持续采集。",{"user":181,"desc":182},"受保护站点","普通请求被封时用StealthyFetcher。",{"user":184,"desc":185},"生产管线","一套代码里快速HTTP+浏览器兜底。",{"user":187,"desc":188},"AI抽取","通过MCP让助手驱动抓取。",{"user":190,"desc":191},"缺维护的团队","避免每次UI改版就重写选择器。",[193,199,202],{"name":194,"price":195,"period":9,"features":196},"开源(BSD-3)","免费",[197,198],"完整库","自托管",{"name":198,"price":195,"period":9,"features":200},[201],"自备基础设施\u002F代理",{"name":203,"price":204,"period":9,"features":205},"可选集成","可变",[206],"按需接验证码\u002F代理服务",[208,211,214,217],{"q":209,"a":210},"Scrapling独特在哪？","自适应解析在网站变化后重定位元素，外加内置隐秘抓取。",{"q":212,"a":213},"自适应怎么实现？","先存结构\u002F文本指纹，选择器失效时按相似度重新匹配。",{"q":215,"a":216},"能过Cloudflare吗？","StealthyFetcher可自动过Turnstile，但没有工具能永远绕过一切。",{"q":218,"a":219},"合法吗？","只采集你有权采集的数据，遵守条款与法律。",[221,246,289,314,341,382,408,430,454,479,503,530],{"slug":222,"name":222,"cnName":223,"category":22,"categories":224,"subcategory":9,"pricing":24,"url":225,"logo":9,"accent":226,"tags":227,"rating":25,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":233,"monthlyVisits":25,"inputs":237,"outputs":238,"en":239,"zh":242,"similarity":245},"curl-impersonate","curl-impersonate 浏览器指纹请求",[22],"https:\u002F\u002Fgithub.com\u002Flwthiker\u002Fcurl-impersonate","#475569",[228,229,44,230,231,232,49],"TLS fingerprint","JA3","curl","browser impersonation","指纹伪装",[29,28,234,235,236],"Node.js","Linux","macOS",[32,33],[35,36,37],{"tagline":240,"description":241},"A special curl build with matching browser TLS\u002FJA3 and HTTP\u002F2 fingerprints, so requests look exactly like real Chrome or Firefox and beat TLS-fingerprint blocking.","curl-impersonate is a fork of curl (and companion wrappers) whose TLS client hello, cipher order, extensions and HTTP\u002F2 settings are modified to be byte-identical to real Chrome, Firefox or Safari. Standard HTTP libraries betray themselves through TLS\u002FJA3 and HTTP2 fingerprints even with a normal user-agent; this tool removes that tell, which is the low-level technique many expensive scraping APIs use under the hood. It runs as a drop-in curl replacement or via Python\u002FNode wrappers, helping requests reach sites that block non-browser TLS signatures.",{"tagline":243,"description":244},"特殊版 curl：TLS\u002FJA3 与 HTTP\u002F2 指纹与真实Chrome\u002FFirefox完全一致，让请求看起来就像真人浏览器，突破TLS指纹封锁。","curl-impersonate 是 curl 的分支（及配套封装），把 TLS Client Hello、加密套件顺序、扩展与 HTTP\u002F2 设置改到与真实 Chrome\u002FFirefox\u002FSafari 逐字节一致。普通HTTP库即使改了 User-Agent，也会在 TLS\u002FJA3 与 HTTP2 指纹上露馅；本工具正是抹掉这一破绽——这也是许多昂贵采集API底层秘而不宣的技巧。它可作为 curl 的直接替代，也有 Python\u002FNode 封装，帮助请求抵达那些按非浏览器TLS签名封人的网站。",38,{"slug":247,"name":248,"cnName":249,"category":22,"categories":250,"subcategory":9,"pricing":24,"url":252,"logo":9,"accent":253,"tags":254,"rating":267,"saves":268,"featured":53,"editorPick":53,"priceFrom":113,"reviewCount":25,"updatedAt":54,"platforms":269,"monthlyVisits":9,"inputs":271,"outputs":272,"siteHealth":275,"en":282,"zh":285,"similarity":288},"mediacrawler","MediaCrawler","MediaCrawler 自媒体爬虫",[22,251],"research","https:\u002F\u002Fgithub.com\u002FNanmiCoder\u002FMediaCrawler","#0EA5E9",[255,256,257,258,259,260,261,46,262,263,264,265,266],"web scraper","crawler","xiaohongshu","douyin","bilibili","weibo","zhihu","social media","爬虫","小红书","抖音","舆情",4.4,7793,[270,236,235,29],"Windows",[33,32],[35,273,274],"CSV","JSON",{"status":276,"http":277,"latencyMs":278,"finalUrl":252,"checkedAt":279,"title":280,"lang":281},"alive",200,559,"2026-09-29T17:30:09Z","GitHub - NanmiCoder\u002FMediaCrawler: 小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B 站视频 ｜ 评论爬虫、微博帖子 ｜ 评论爬虫、百度贴吧帖子 ｜ 百度贴吧评论回复爬虫 | 知乎","en",{"tagline":283,"description":284},"An open-source Python crawler for seven major Chinese social platforms — Xiaohongshu, Douyin, Kuaishou, Bilibili, Weibo, Tieba and Zhihu — driven by a real Playwright browser so no signature reverse-engineering is needed.","MediaCrawler (by NanmiCoder, ~60k GitHub stars) is a widely used self-media data-collection framework for the Chinese internet. It covers seven platforms — Xiaohongshu (RED), Douyin, Kuaishou, Bilibili, Weibo, Baidu Tieba and Zhihu — and can gather posts, videos, notes, comments (including second-level replies), creator-homepage content and search results. Instead of manually reverse-engineering each platform's encrypted signatures, it uses Playwright to drive a real browser, keeps the logged-in context (QR-code or cookie login, cached for reuse) and obtains encryption parameters by evaluating JavaScript, which greatly lowers the barrier. Across platforms it supports keyword search, crawl-by-post-ID, creator-homepage crawl, login-state caching, an IP proxy pool and comment word-cloud generation, with results saved to CSV\u002FJSON or MySQL\u002FPostgreSQL\u002FSQLite. Setup uses Python 3.11 (uv recommended), Node.js 16+ for some platforms and Playwright browsers. The project is open for learning and research on public data only; commercial use requires authorization, and users must respect each platform's terms, data-protection law and account-safety rules.",{"tagline":286,"description":287},"开源的中文自媒体爬虫框架，一套代码覆盖小红书、抖音、快手、B站、微博、贴吧、知乎7大平台，用 Playwright 驱动真实浏览器，免去手写加密签名逆向。","MediaCrawler（作者 NanmiCoder，GitHub 约6万 star）是国内使用极广的自媒体数据采集框架，覆盖小红书、抖音、快手、B站、微博、百度贴吧、知乎7大平台，可采集笔记\u002F视频\u002F帖子、评论（含二级评论回复）、创作者主页内容与搜索结果。它不手工逆向各平台加密签名，而是用 Playwright 驱动真实浏览器、保留登录上下文（扫码或Cookie登录并缓存复用），通过执行JS表达式获取加密参数，大幅降低逆向门槛。各平台普遍支持关键词搜索、按帖子\u002F视频ID采集、创作者主页采集、登录态缓存、IP代理池、评论词云图，结果可存为 CSV\u002FJSON 或 MySQL\u002FPostgreSQL\u002FSQLite。环境建议 Python 3.11（推荐 uv 管理依赖）、部分平台需 Node.js 16+ 并安装 Playwright 浏览器。项目仅面向公开数据的学习与研究，商用需单独授权，使用时须遵守各平台条款、数据合规法律与账号安全规则。",35,{"slug":290,"name":291,"cnName":292,"category":22,"categories":293,"subcategory":9,"pricing":24,"url":294,"logo":9,"accent":295,"tags":296,"rating":25,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":303,"monthlyVisits":25,"inputs":306,"outputs":307,"en":308,"zh":311,"similarity":288},"crawlee","Crawlee","Crawlee 生产级爬虫框架",[22],"https:\u002F\u002Fcrawlee.dev","#7C3AED",[297,298,299,46,300,301,302],"scraping framework","proxy rotation","headless browser","puppeteer","爬虫框架","防封",[304,305,28,29],"JavaScript","TypeScript",[32,33],[35,36,37],{"tagline":309,"description":310},"A production-grade web-scraping and browser-automation library for JavaScript\u002FTypeScript and Python, by Apify — queues, retries, proxy rotation and fingerprints built in.","Crawlee is a free, Apache-2.0 scraping framework (by Apify) that handles the hard parts of crawling at scale: persistent request queues (breadth\u002Fdepth first), auto-scaling concurrency, automatic retries and error handling, integrated proxy rotation and session pooling, and browser-fingerprint management so crawlers look human. It offers one interface across fast HTTP\u002FCheerio crawling and headless Puppeteer\u002FPlaywright crawling, plus pluggable datasets, key-value stores, Dockerfiles and a CLI to bootstrap projects — scaling from a single page to millions of URLs without swapping tools.",{"tagline":312,"description":313},"Apify 出品的生产级网页抓取与浏览器自动化库（JS\u002FTS 与 Python）：队列、重试、代理轮换、浏览器指纹一应俱全。","Crawlee 是 Apify 出品的免费 Apache-2.0 抓取框架，专门解决大规模爬取的难题：持久化请求队列（广度\u002F深度优先）、按资源自动扩缩并发、自动重试与错误处理、内置代理轮换与会话池、浏览器指纹管理让爬虫更像真人。它用统一接口同时覆盖快速的 HTTP\u002FCheerio 抓取与无头 Puppeteer\u002FPlaywright 抓取，并自带数据集、键值存储、Dockerfile 与项目脚手架CLI，从单页到百万URL无需换工具。",{"slug":315,"name":316,"cnName":317,"category":22,"categories":318,"subcategory":319,"pricing":24,"url":320,"logo":9,"accent":253,"tags":321,"rating":25,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":328,"monthlyVisits":25,"inputs":330,"outputs":331,"en":334,"zh":337,"similarity":340},"dokploy","Dokploy","Dokploy 自托管部署平台",[22],"dev-platforms","https:\u002F\u002Fdokploy.com",[322,323,324,325,326,47,198,327],"self-hosted","paas","docker swarm","git deploy","deployment","一键部署",[235,236,270,329,29],"Web",[37,32],[332,333],"Deployment","Web App",{"tagline":335,"description":336},"A free, self-hostable PaaS powered by Docker Swarm: connect a Git repo and auto-deploy on push, with one-command install, free SSL, managed databases, real-time monitoring and AI-assisted deployments.","Dokploy is an open-source, self-hosted deployment platform (a Vercel\u002FHeroku alternative) installed on a Linux server with a single command and powered by Docker Swarm. It connects to GitHub, GitLab, Gitea, Bitbucket, a Docker registry or generic Git, and deploys automatically on every push via webhooks, with monorepo support. Beyond applications it manages domains and free SSL certificates, managed databases, one-click apps and services, environment variables, and scheduled backups. A real-time dashboard tracks CPU, memory and network with alerts, and newer releases add AI-assisted deployment through MCP plus governed sandboxes with SSO and multi-tenancy. It aims to give teams a clean, low-cost control plane on infrastructure they own.",{"tagline":338,"description":339},"免费、可自托管的PaaS，基于 Docker Swarm：连接Git仓库即推送自动部署，一条命令安装，含免费SSL、托管数据库、实时监控与AI辅助部署。","Dokploy 是开源、可自托管的部署平台（Vercel\u002FHeroku替代品），在Linux服务器上一条命令即可安装，底层用 Docker Swarm。它连接 GitHub\u002FGitLab\u002FGitea\u002FBitbucket、Docker镜像仓库或通用Git，通过webhook在每次推送时自动部署，并支持monorepo。除应用外，还管理域名与免费SSL证书、托管数据库、一键应用与服务、环境变量、定时备份。实时面板监控CPU\u002F内存\u002F网络并告警，新版本还通过 MCP 提供AI辅助部署，以及带SSO与多租户的受控沙箱。目标是在你自有的基础设施上，提供干净、低成本的部署控制面。",34,{"slug":342,"name":343,"cnName":343,"category":22,"categories":344,"subcategory":319,"pricing":24,"url":345,"accent":26,"tags":346,"rating":360,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":361,"verified":53,"platforms":362,"inputs":367,"outputs":372,"en":375,"zh":378,"similarity":381},"rea-reverse-engineer","REA",[22,251],"https:\u002F\u002Frea.tools",[347,348,349,47,350,351,352,353,354,355,356,357,358,359],"reverse engineering","decompiler","disassembler","binary analysis","static analysis","Ghidra","Hopper","IDA","CTF","security research","open source","developer tools","agent skills",4.8,"2026-10-11",[29,47,363,364,365,366],"Claude Code","Codex","Cursor","Gemini CLI",[368,369,370,371],"file","binary","app","url",[22,373,374],"analysis","report",{"tagline":376,"description":377},"Reverse engineer anything with agents — from app behavior down to native binaries.","REA (Reverse Engineer Anything) is a free, open-source (MIT) MCP server that lets AI agents inspect applications without their source code. It connects Claude Code, Codex, Cursor, Gemini CLI and other agents to tools for analyzing native binaries, JavaScript and Electron apps, .NET assemblies, Android APKs, websites, firmware and more — returning pseudocode, assembly, modules, imports, evidence and limitations behind each conclusion. Ask your agent to understand how a feature in any app works, show the evidence, and build a similar version for your own project. Deep native analysis uses Hopper, Ghidra or IDA; static JavaScript and .NET inspection work without a native engine. All analysis runs locally (nothing is uploaded) and the same tools are available from the terminal via the rea CLI. 65.2K GitHub stars. Setup: npx rea-agents setup; CLI: npm install --global rea-agents.",{"tagline":379,"description":380},"用 Agent 逆向一切——从应用行为到原生二进制。","REA（Reverse Engineer Anything）是一个免费开源的（MIT）MCP 服务器，让 AI 代理在没有源代码的情况下检查应用程序。它把 Claude Code、Codex、Cursor、Gemini CLI 等代理连接到分析工具，可分析原生二进制、JavaScript 和 Electron 应用、.NET 程序集、Android APK、网站、固件等，并返回伪代码、汇编、模块、导入关系以及每条结论的证据和局限。你可以让代理理解任何应用中的某个功能是如何工作的、展示证据，并为自己的项目重建类似功能。深度原生分析使用 Hopper、Ghidra 或 IDA；静态 JavaScript 和 .NET 检查无需原生引擎。所有分析都在本地进行（不上传任何内容），同样的工具也可通过 rea CLI 在终端使用。GitHub 65.2K stars。安装：npx rea-agents setup；CLI：npm install --global rea-agents。",33,{"slug":383,"name":384,"cnName":385,"category":22,"categories":386,"subcategory":9,"pricing":24,"url":387,"logo":9,"accent":388,"tags":389,"rating":25,"saves":25,"featured":53,"editorPick":397,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":398,"monthlyVisits":25,"inputs":399,"outputs":400,"en":401,"zh":404,"similarity":407},"crawl4ai","Crawl4AI","Crawl4AI 大模型爬虫",[22,251],"https:\u002F\u002Fgithub.com\u002Funclecode\u002Fcrawl4ai","#22C55E",[256,390,391,392,393,394,395,396],"LLM","async","markdown","RAG","no api key","开源爬虫","大模型",true,[329,29,30,28],[32,33],[35,36,37],{"tagline":402,"description":403},"A free, open-source crawler built for LLMs: async, extremely fast, turns any site into clean markdown or structured JSON locally, with no API key or account.","Crawl4AI is an open-source Python crawler created specifically to feed large language models. It drives a browser asynchronously to render JavaScript-heavy pages and converts them into clean, LLM-friendly markdown or structured JSON, with configurable extraction strategies, chunking, metadata and screenshots. Because it runs on your own machine there is no API key, no account and no per-page fee — a point its author famously made after being frustrated by paid scraping services. It integrates with LangChain and similar frameworks and is optimized for high-throughput, cost-free local pipelines.",{"tagline":405,"description":406},"免费开源、为大模型而生的爬虫：异步极速，本地把任意网站转成干净 Markdown 或结构化 JSON，无需 API Key、无需账户。","Crawl4AI 是专为喂养大模型而做的开源 Python 爬虫。它异步驱动浏览器渲染 JS 重页面，转成干净、对 LLM 友好的 Markdown 或结构化 JSON，并提供可配置的抽取策略、分块、元数据与截图。因为完全在本机运行，所以无需 API Key、账户或按页付费——作者正是被付费爬虫服务激怒后几天内做出了它。它可与 LangChain 等框架集成，面向高吞吐、零成本的本地数据管线优化。",32,{"slug":409,"name":410,"cnName":411,"category":22,"categories":412,"subcategory":9,"pricing":24,"url":413,"logo":9,"accent":414,"tags":415,"rating":25,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":421,"monthlyVisits":25,"inputs":422,"outputs":423,"en":424,"zh":427,"similarity":407},"scrapy","Scrapy","Scrapy 工业级爬虫框架",[22],"https:\u002F\u002Fscrapy.org","#A32020",[416,417,418,419,391,301,420],"crawling framework","spider","middleware","pipelines","数据采集",[28,29],[32,33],[35,36,37],{"tagline":425,"description":426},"The battle-tested open-source Python crawling framework powering data teams for 15+ years: fast async spiders, middleware, pipelines and exports at massive scale.","Scrapy is the classic industrial-strength web-crawling framework for Python. A project is built from Spiders that emit requests and parse responses, with an async engine that fetches many pages concurrently, a middleware layer for cookies, headers, user-agent spoofing, retries, caching and robots.txt, and item pipelines that clean, validate and export data to JSON, CSV, XML or databases. Huge extension ecosystem (e.g. scrapy-playwright for JavaScript pages), built-in caching, throttling and telnet console make it the standard choice when reliability and scale matter more than a quick one-off script.",{"tagline":428,"description":429},"15年以上久经考验的开源Python爬虫框架：高速异步蜘蛛、中间件、数据管道与大规模导出，数据团队的工业级标准。","Scrapy 是经典的工业级 Python 网络爬取框架。一个工程由若干“蜘蛛(Spider)”组成：发出请求、解析响应；异步引擎高并发抓取大量页面；中间件层处理 Cookie、请求头、UA伪装、重试、缓存与robots；数据管道负责清洗、校验并把条目导出为 JSON\u002FCSV\u002FXML 或写入数据库。庞大的扩展生态（如 scrapy-playwright 支持JS页面）、内置缓存、限速与Telnet控制台，使其成为“要可靠、要规模、而非一次性脚本”时的标准选择。",{"slug":431,"name":432,"cnName":433,"category":22,"categories":434,"subcategory":9,"pricing":24,"url":436,"logo":9,"accent":437,"tags":438,"rating":25,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":445,"monthlyVisits":25,"inputs":446,"outputs":447,"en":448,"zh":451,"similarity":407},"markitdown","MarkItDown","微软 MarkItDown 文档转Markdown",[22,435],"productivity","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fmarkitdown","#0F6CBD",[392,439,440,441,442,393,390,443,444],"document conversion","pdf","office","ocr","格式转换","微软",[29,28,30],[32,33],[35,36,37],{"tagline":449,"description":450},"Microsoft's open-source tool that converts almost anything — PDF, Office docs, images (OCR), HTML, audio, archives and web pages — into clean Markdown for LLMs.","MarkItDown is a lightweight Microsoft open-source library\u002FCLI that transforms a very wide range of files into uniform Markdown so they can be fed to language models: Word, PowerPoint, Excel, PDF, HTML, CSV\u002FJSON\u002FXML, images (with OCR\u002Fdescriptions), audio (with transcription), EPUB, ZIP archives and more, plus web-page URLs. It optionally embeds metadata and image descriptions, is pip-installable and scriptable in Python, and has become a common first stage in RAG and document-AI pipelines where messy source formats must become clean text.",{"tagline":452,"description":453},"微软开源工具，把几乎一切——PDF、Office文档、图片(OCR)、HTML、音频、压缩包和网页——统一转成干净 Markdown 喂给大模型。","MarkItDown 是微软开源的轻量库\u002F命令行工具，能把极多种文件统一转成 Markdown 以便喂给语言模型：Word、PowerPoint、Excel、PDF、HTML、CSV\u002FJSON\u002FXML、图片（OCR\u002F描述）、音频（转写）、EPUB、ZIP 压缩包乃至网页URL。它可选择性保留元数据与图片描述，pip 即可安装、可用 Python 调用，已成为 RAG 与文档AI管线中“把杂乱源格式变成干净文本”的常见第一步。",{"slug":455,"name":455,"cnName":456,"category":22,"categories":457,"subcategory":9,"pricing":24,"url":458,"logo":9,"accent":459,"tags":460,"rating":25,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":469,"monthlyVisits":25,"inputs":471,"outputs":472,"en":473,"zh":476,"similarity":407},"scrcpy","scrcpy 安卓投屏控制",[22,435],"https:\u002F\u002Fgithub.com\u002FGenymobile\u002Fscrcpy","#16A34A",[461,462,463,464,465,466,467,468],"android","screen mirror","adb","remote control","mobile bridge","no root","安卓投屏","移动采集",[270,236,235,470],"Android",[32,33],[35,36,37],{"tagline":474,"description":475},"Genymobile's open-source tool to display and fully control Android devices from a computer over USB or Wi-Fi — no root, low latency, with audio and recording.","scrcpy (Genymobile\u002FGenymotion) mirrors and lets you control Android phones\u002Ftablets from a Windows\u002FMac\u002FLinux computer over USB or TCP, with no root access and very low latency by sending H.264\u002FH.265 video from the device. It supports keyboard\u002Fmouse input, clipboard sync, file drag-and-drop, screen recording, Android 11+ audio forwarding, multi-device and wireless (adb) connections. For data work it is a bridge to mobile-only apps and content that ordinary web crawlers cannot reach, letting you automate or observe the Android UI on a desktop.",{"tagline":477,"description":478},"Genymobile 开源工具：通过USB或Wi-Fi在电脑上显示并完整操控安卓设备，无需root、低延迟，支持音频与录屏。","scrcpy（Genymobile\u002FGenymotion）可在 Windows\u002FMac\u002FLinux 上通过 USB 或 TCP 镜像并操控安卓手机\u002F平板，无需 root，借助设备端 H.264\u002FH.265 视频流实现极低延迟。支持键鼠输入、剪贴板同步、文件拖拽、屏幕录制、Android 11+ 音频转发、多设备与无线(adb)连接。在数据采集中，它是通往“纯移动App与内容”的桥梁——普通网页爬虫够不到的移动端，可在桌面上观察或自动化安卓界面。",{"slug":480,"name":481,"cnName":482,"category":22,"categories":483,"subcategory":9,"pricing":24,"url":484,"logo":9,"accent":485,"tags":486,"rating":25,"saves":25,"featured":53,"editorPick":53,"priceFrom":25,"reviewCount":25,"updatedAt":54,"verified":53,"platforms":494,"monthlyVisits":25,"inputs":495,"outputs":496,"en":497,"zh":500,"similarity":407},"autoscraper","AutoScraper","AutoScraper 示例智能爬虫",[22],"https:\u002F\u002Fgithub.com\u002Falirezamika\u002Fautoscraper","#0891B2",[487,488,489,490,491,492,493],"smart scraper","example based","no selectors","beginner","rapid prototype","智能爬虫","免选择器",[28,29],[32,33],[35,36,37],{"tagline":498,"description":499},"A 'smart' Python scraper: show it one example result and it learns the pattern to fetch every similar item — no CSS\u002FXPath selectors and almost no code.","AutoScraper is a lightweight, beginner-friendly Python library that learns scraping rules from examples. You give it a URL and one sample of the text you want (a title, a price, a link); it infers the matching rules and returns every similar item on the page, and the learned model can be reused on other pages of the same site. This removes the need to inspect and maintain XPath or CSS selectors for straightforward, repetitive listings, making it a fast 'give me the data' tool for prototypes and small jobs.",{"tagline":501,"description":502},"“聪明”的Python爬虫：给它一个示例结果，它就学会规律并抓出所有同类内容——无需CSS\u002FXPath选择器，几行代码搞定。","AutoScraper 是轻量、对新手友好的 Python 库，用“示例”学习抓取规则：你给它一个网址和一个你想要的样本文本（标题、价格、链接），它推断匹配规则并返回页面上所有同类项，学到的模型还能在同站其他页面复用。对简单重复的列表页，免去查看和维护 XPath\u002FCSS 选择器，是原型和小任务里“直接给我数据”的快捷工具。",{"slug":504,"name":505,"category":22,"categories":506,"subcategory":319,"pricing":24,"url":507,"logo":9,"accent":508,"tags":509,"rating":512,"saves":513,"editorPick":53,"priceFrom":113,"reviewCount":25,"updatedAt":9,"platforms":514,"monthlyVisits":9,"inputs":517,"outputs":518,"siteHealth":519,"en":523,"zh":526,"similarity":529},"github-32","PGVector",[],"https:\u002F\u002Fgithub.com\u002Fpgvector\u002Fpgvector","#9C27B0",[510,511],"AI Code","Developer Tools",4.1,9918,[329,270,515,235,516,470],"Mac","iOS",[33,37],[37,33],{"status":276,"http":277,"latencyMs":520,"finalUrl":507,"checkedAt":521,"title":522,"lang":281},664,"2026-09-29T09:45:15Z","GitHub - pgvector\u002Fpgvector: Open-source vector similarity search for Postgres",{"tagline":524,"description":525},"Postgres vector extension.","Open-source vector similarity search for Postgres. Contribute to pgvector\u002Fpgvector development by creating an account on GitHub.",{"tagline":527,"description":528},"AI驱动的code工具","PGVector 是一个 PostgreSQL 扩展，用于存储和查询向量数据。它支持精确和近似最近邻搜索，并集成了多种索引类型，如 HNSW 和 IVFFlat。该扩展易于安装，无需额外服务，可直接在现有 PostgreSQL 数据库中使用。适用于推荐系统、图像检索和自然语言处理等场景。",31,{"slug":531,"name":532,"category":22,"categories":533,"subcategory":534,"pricing":24,"url":535,"logo":9,"accent":536,"tags":537,"rating":539,"saves":540,"editorPick":53,"priceFrom":113,"reviewCount":25,"updatedAt":9,"platforms":541,"monthlyVisits":9,"inputs":543,"outputs":544,"siteHealth":545,"en":555,"zh":557,"similarity":529},"code-genius","Code Genius",[],"code-assistant","https:\u002F\u002Fwww.code-genius.dev","#00E676",[510,511,538],"Trending",4.7,9905,[329,542],"API",[33,37],[37,33],{"status":546,"http":25,"latencyMs":25,"finalUrl":535,"checkedAt":547,"attempts":548},"error","2026-09-29T09:45:18Z",[549,552,553,554],[550,551],"http",301,[550,551],[550,551],[550,551],{"tagline":556,"description":556},"Streamlines React, Vue JS, and Tailwind CSS development.",{"tagline":527,"description":558},"Code Genius 是一款 AI 驱动的代码工具，专注于简化 React、Vue.js 和 Tailwind CSS 的开发流程。它提供智能代码补全、实时错误检测和代码解释功能，支持 Python、JavaScript、TypeScript 等多种编程语言，帮助开发者提高编码效率，快速理解和修复代码问题。目前免费使用，是开发者的得力助手。"]