待翻譯：Show HN: OptiQra – Open-source AI website optimization and intelligence platform

AI 服務暫時不可用，以下為來源正文，待恢復後補全翻譯。

Notifications You must be signed in to change notification settings Fork 0 Star 2 BranchesTags Open more actions menu Folders and files NameName Last commit message Last commit date Latest commit History 228 Commits 228 Commits .github .github .next-tauri-placeholder .next-tauri-placeholder dist-server dist-server public public server server src-tauri src-tauri src src tests/e2e tests/e2e worker worker .dockerignore .dockerignore .editorconfig .editorconfig .env.example .env.example .gitignore .gitignore .nvmrc .nvmrc AGENTS.md AGENTS.md CLAUDE.md CLAUDE.md CONTRIBUTING.md CONTRIBUTING.md DEPLOYMENT.md DEPLOYMENT.md Dockerfile Dockerfile Dockerfile.dev Dockerfile.dev LICENSE LICENSE README.md README.md SECURITY.md SECURITY.md desktop-icon.png desktop-icon.png docker-compose.dev.yml docker-compose.dev.yml docker-compose.yml docker-compose.yml eslint.config.mjs eslint.config.mjs next-env.d.ts next-env.d.ts next.config.ts next.config.ts optiqra.webp optiqra.webp package-lock.json package-lock.json package.json package.json playwright.config.ts playwright.config.ts postcss.config.mjs postcss.config.mjs sentry.client.config.ts sentry.client.config.ts sentry.edge.config.ts sentry.edge.config.ts sentry.server.config.ts sentry.server.config.ts showcase.gif showcase.gif tsconfig.json tsconfig.json tsconfig.tsbuildinfo tsconfig.tsbuildinfo Repository files navigation Audit your site for both search engines and AI. OptiQra crawls your entire website and scores it across SEO, performance, accessibility, security, and conversion signals — plus GEO (generative engine optimization) and AEO (answer engine optimization). ChatGPT, Claude, Perplexity, and Google's AI Overviews now send meaningful traffic of their own. OptiQra tells you if your site is set up to be crawled, cited, and answered by them — and then fixes it for you. Live demo · Free, no signup, no account · Also available as a native desktop app for background scans and fully offline project audits Contents Core audits Beyond audits: fixes, not just findings OptiQra Desktop How it works How OptiQra compares Tech stack Quick start Building the desktop app Docker Periodic scans Localization API Project structure Security Roadmap Contributing Core audits 🔍 Site crawling & indexing Crawls entire sites through sitemaps and internal links (configurable page cap) Detects duplicate and near-duplicate content (repeated titles, byte-identical pages, thin/templated content) Analyzes links (broken links, internal/external, redirect chains) Scans images (missing alt text, lazy loading, responsive srcset/ usage) JS-rendered content detection — runs a page's own JavaScript in a sandboxed environment to see the post-hydration DOM, the same way a JS-executing crawler like Googlebot would, instead of guessing from raw HTML alone 🎯 SEO essentials Meta tags (title, description, canonical, Open Graph, Twitter cards) Structured data (JSON-LD, schema.org validation) robots.txt and sitemap analysis, with one-click generation of a starter sitemap/robots.txt when either is missing Internal linking patterns and anchor text quality PageSpeed Insights integration for real Core Web Vitals data 🤖 Generative Engine Optimization (GEO) Entity grounding — checks for authoritative sameAs links (Wikipedia, Wikidata, Crunchbase, LinkedIn, GitHub) that help AI models cite and disambiguate your content Client-side rendering detection — flags content hidden behind client-side JavaScript that generative crawlers (GPTBot, ClaudeBot) can't see Citation-friendliness — analyzes whether pages are structured to be easily quoted and attributed in AI answers llms.txt generator — drafts a starter llms.txt for your site from an AI provider of your choice 🤖 Answer Engine Optimization (AEO) Crawler access — checks whether AI answer-engine crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, etc.) are allowed or blocked by robots.txt Answer-liftable content — detects whether key content is structured to be extracted directly into answers, Q&A format, and featured snippets Q&A structuring — validates whether content is formatted to answer common questions about your domain ⚡ Performance HTML/response performance characteristics Optional PageSpeed Insights scoring for lab + field data ♿ Accessibility Missing alt text, labels, and ARIA attributes Color contrast problems Keyboard navigation and semantic HTML issues 🔐 Security Security headers (HSTS, CSP, X-Frame-Options, etc.) Conversion signals and trust badges Tech-stack fingerprinting (CMS, framework, e-commerce platform, page builder, server/language) so fixes can be written for your actual stack Beyond audits: fixes, not just findings OptiQra doesn't stop at a scored report. It can fix what it finds: Deterministic auto-fix — mechanical issues (a missing tag, a bad attribute, content that already exists elsewhere on the page) are fixed directly, no AI involved, 100% reproducible AI-authored fixes — content that needs actual writing (titles, meta descriptions, alt text, form labels, CTA copy) is batched into a single AI call using your own API key Full AI Fix — an opt-in, deeper pass that hands the model an entire file plus its detected issues and asks for the whole file back, fixed, for anything the deterministic and batched passes couldn't resolve Framework-aware fixes for HTML, JSX/TSX, Vue, Nuxt, Angular, Svelte, and Astro — not just generic markup edits Project (codebase) audits — upload a .zip of your actual project instead of only scanning a live URL, and get the same scored SEO/performance/accessibility categories, plus config-file-level fixes (headers, compression, sitemap/robots.txt) that a live-URL scan can never make since it can't see your source Integrity-checked — every text-splicing fix pass is verified against structural invariants afterward; if anything looks off, the change is discarded rather than shipped A visual "highlighted clone" of any scanned page, annotated element-by-element with the exact issues found, rendered in an iframe instead of only listed in a table Propose upstream — turn a custom rule or settings preset you built into a real GitHub pull request against this repo, via GitHub's own "create file" flow — no token, no server in the middle 🌌 3D / Galaxy crawl visualization Beyond the standard tree view, OptiQra can render a crawled site as an interactive 3D structure or an animated "galaxy" of pages (Three.js), letting you spot structural and performance patterns across hundreds of pages at a glance. 🎭 AI insights, moods, and custom rules Site-wide AI insights — a strategic summary reasoning across every category and page AI moods & potency — pick the persona the AI writes fixes/insights in, and how strongly it leans into that persona Custom rules — small, client-side JS snippets you author to post-process a finished scan and surface additional findings specific to your site Custom code — inject raw CSS or a one-off JS snippet into the report view itself 📊 Reports & data Multi-provider AI — paste an API key (yours to keep) for OpenAI, Anthropic, Google, Groq, OpenRouter, Mistral, DeepSeek, or xAI (Grok) Export to PDF, DOCX, XLSX, Markdown, CSV, TSV, TXT, JSON 13 languages — the entire UI is translatable (see Localization) Periodic scanning — schedule re-scans (hourly to yearly), get notifications, auto-compare against previous results PWA + offline — installable in-browser, past reports visible without internet 🖥️ OptiQra Desktop OptiQra also ships as a native desktop app, built with Tauri 2 (Rust shell + the same Next.js engine as the web app, running locally). It exists to solve one specific limitation of the browser version: A browser-based scheduler only fires while a tab is open. The desktop app fires schedules from a background process that keeps running after you close the window. What's actually different, not just wrapped: Web app Desktop app Scheduled scans Only while a tab/PWA is open Run from a background daemon — the window can be fully closed Project (codebase) audits Needs a network round-trip Fully offline — the audit engine runs on your machine Live-site crawling & AI fixes Needs a network round-trip Still needs a network round-trip (crawling a live site or calling an AI provider inherently does) Notifications Browser Notification API Real OS notifications, even with no window visible Scores/ruleset Same engine Identical engine — no separate desktop ruleset to drift out of sync How it works under the hood: The Rust shell spawns the same Next.js app (built output: "standalone" and packaged as a single executable) as a background sidecar process, and points the window's webview at it — it does not reimplement the crawler/audit/fix engine in Rust. Closing the window hides it instead of quitting; only the tray menu's "Quit" actually stops the sidecar (and therefore the scheduler). A scheduler daemon inside the sidecar mirrors the same due/compare/notify logic the browser's scheduler.ts uses, so a schedule behaves identically whether a tab or the daemon catches it — it just reads/writes a local file store instead of IndexedDB, and notifies through the OS instead of the browser. The browser UI's schedule/scan stores and the daemon's file store are kept in sync automatically: creating or editing a schedule mirrors into the daemon's store, and any scan the daemon ran while the window was closed appears in Recent Scans the next time you open the app. Single-instance locking, autostart, and an in-app updater (checked against GitHub Releases) are included. Platforms: macOS 12+ (Apple Silicon & Intel), Windows 10/11 (x64), Linux (AppImage / .deb). Download the latest build from the releases page, or see Building the desktop app to build it yourself. How it works Paste a URL (or upload a project .zip) — any website, any size OptiQra crawls it — follows internal links, scans every page (or every file, in project mode) See results in seconds — visual/3D/galaxy tree of all crawled pages, scores by category, detailed issue list Apply fixes — deterministic fixes apply automatically; content-authoring fixes use your chosen AI provider Get AI insights — (optional) a site-wide strategic summary reasoning across all categories and pages Everything runs in your browser (or, with the desktop app, entirely on your machine). No server sees your data unless you're scanning a live URL. No account needed. How OptiQra compares Feature OptiQra Lighthouse SEMrush Ahrefs Full-site crawl ✅ Free ❌ Single page ✅ Paid ✅ Paid GEO/AEO audits ✅ Yes ❌ No ❌ No ❌ No AI-generated & applied fixes ✅ Yes (your API key) ❌ No ❌ No ❌ No Codebase / project audits ✅ Yes, offline in desktop app ❌ No ❌ No ❌ No Background scheduled scans ✅ Yes (desktop app) ❌ No ✅ Paid ✅ Paid No signup required ✅ Yes ✅ Yes ❌ No ❌ No Runs in browser / offline ✅ Yes ✅ Yes ❌ No ❌ No Free forever ✅ Yes ✅ Yes ❌ No ❌ No Open source ✅ Yes ✅ Yes ❌ No ❌ No OptiQra's edge: free full-site crawls, AI-native audits (GEO/AEO), fixes that actually apply (not just suggestions), and a desktop app for offline project audits and background scheduling — all with no account and no subscription. Why this matters SEO isn't just about Google anymore. ChatGPT, Claude, Perplexity, and Gemini now generate answers that cite (or don't cite) your site. But most auditing tools haven't caught up — they still measure you against Googlebot alone. OptiQra fills that gap. It tells you: Whether your content is visible to AI crawlers (or hidden behind client-side JS) Whether AI models can cite yo [truncated for AI cost control]