GEO Audit for wolframalpha.com

{"domain": "wolframalpha.com", "overall_score": 57, "business_model": "Freemium Computational Knowledge Engine (B2C consumers + B2B API/LLM integrations)", "audience_tags": ["To-C Students & Educators", "To-C STEM Professionals", "To-B Developers (API)", "To-B Enterprise (LLM/agent integration)"], "tech_stack": ["Next.js", "styled-components", "Wolfram Engine (backend computation)", "CDN edge redirects (apex->www)"], "competitor_domains": ["symbolab.com", "mathway.com", "desmos.com", "wolfram.com"], "scorecard": [{"metric": "Technical Accessibility", "score": 10, "max": 15, "status": "WARNING", "detail": "No crawl barriers; SSR 200 for all crawlers. robots.txt wildcard only disallows /facebook/. OAI-SearchBot, PerplexityBot, ClaudeBot allowed via wildcard. However GPTBot, ChatGPT-User and Google-Extended are explicitly disallowed at /, which limits OpenAI training-data ingestion and some ChatGPT-referenced surfaces."}, {"metric": "Structured Schema Quality", "score": 1, "max": 10, "status": "ACTION_REQUIRED", "detail": "Zero JSON-LD blocks across all probed pages (/, /examples/math, /about, /pro). No Organization, WebSite, SearchAction, SoftwareApplication or FAQPage schema anywhere. Major gap for a market-leading knowledge engine."}, {"metric": "Content Extractability", "score": 9, "max": 15, "status": "WARNING", "detail": "Next.js SSR delivers ~13KB of real body text: clear value proposition and 4 H2 category sections (Mathematics, Science & Technology, Society & Culture, Everyday Life). But no H1 and no canonical tag; heading hierarchy is thin."}, {"metric": "Answer Density (BLUF)", "score": 7, "max": 10, "status": "PASSED", "detail": "Meta description and hero copy give a crisp bottom-line: 'Compute expert-level answers using Wolfram's breakthrough algorithms, knowledgebase and AI technology'. Not formatted as an explicit 40-70 word definition block but easily extractable."}, {"metric": "Citations & Empirical Data", "score": 6, "max": 15, "status": "WARNING", "detail": "Homepage marketing copy contains almost no hard numbers ('relied on by millions' is the only quantitative claim). No benchmarks, data citations, or verifiable stats on landing pages; empirical strength lives inside query results (dynamic), not in crawlable static content."}, {"metric": "Expertise & Attribution (EEAT)", "score": 8, "max": 10, "status": "PASSED", "detail": "Wolfram Research + Stephen Wolfram carry very strong brand authority; /about exists; deep academic/technical ecosystem (reference.wolfram.com, Wolfram Language). No per-page bylines needed for a company brand of this stature."}, {"metric": "Off-Page Entity Footprint", "score": 12, "max": 15, "status": "PASSED", "detail": "Wikipedia entry present; WolframResearch GitHub org (59 public repos, WolframLanguageForJupyter 1212 stars); active Twitter @Wolfram_Alpha; extensive media/academic mentions since 2009."}, {"metric": "Machine-Readable Discovery", "score": 4, "max": 10, "status": "WARNING", "detail": "OG tags and Twitter Cards are complete. But sitemap.xml returns 404 on www (apex 301s to www then 404s); no llms.txt, no skill.md, no agents.md, no .well-known/mcp/server-card.json, no canonical, no meta robots."}], "prompt_fan_out": ["what is the population of France in 2024", "solve the integral of x^2 * sin(x)", "difference between WolframAlpha and ChatGPT for math problems", "how many calories in a banana", "compute the derivative of e^x divided by x^2 + 1"], "recommendations": ["Fix sitemap.xml: /sitemap.xml currently 404s on www.wolframalpha.com after the apex 301 redirect — serve a real XML sitemap (index + children) so crawlers can discover the full URL inventory.", "Add JSON-LD structured data sitewide: Organization + WebSite (with SearchAction for the query endpoint), SoftwareApplication for Pro, and FAQPage on high-traffic examples pages. Currently zero schema markup exists.", "Add canonical tags on every page (currently absent) and keep the apex->www 301 canonicalization consistent.", "Publish agent-native discovery files at the root: llms.txt (with pointers to API docs and examples index), skill.md, and agents.md to make the engine discoverable to AI agents.", "Reconsider the GPTBot/ChatGPT-User/Google-Extended blanket disallow — at minimum allow OAI-SearchBot explicitly (it currently falls through to the wildcard) so ChatGPT Search surfaces fresh query results, while keeping training-crawler blocks if desired.", "Add a proper H1 and a 40-70 word BLUF summary block under it on the homepage to tighten machine extraction.", "Add concrete empirical data points to marketing pages (curated data domains count, API latency/uptime, steps solved) so LLMs can cite verifiable numbers."], "agent_repair_prompt": "You are an AI coding agent tasked with fixing GEO (Generative Engine Optimization) gaps on wolframalpha.com, a Next.js (App/Pages Router, styled-components) site. Implement these fixes and verify each: (1) Add a functional sitemap: create app/sitemap.ts or pages/sitemap.xml generating a valid XML sitemap at /sitemap.xml referencing core routes (/, /examples, /examples/math, /about, /pro) — confirm with `curl -sL https://www.wolframalpha.com/sitemap.xml | head -c 100` returns `<?xml`. (2) Inject JSON-LD: add a Script type=\"application/ld+json\" with an Organization schema (name: Wolfram|Alpha, url: https://www.wolframalpha.com, sameAs: [Wikipedia, Twitter]) and a WebSite schema with potentialAction SearchAction targeting the query URL pattern, on the homepage; add FAQPage schema to at least /examples/math. (3) Add rel=canonical to every page pointing to the www canonical URL. (4) Create llms.txt at the root describing the service, key query URL format (https://www.wolframalpha.com/input?i=<query>), API docs location, and examples index; create a short skill.md and agents.md. (5) Update robots.txt: keep training crawler blocks if desired but add explicit `User-agent: OAI-SearchBot / Allow: /` and `User-agent: PerplexityBot / Allow: /` groups so AI search engines can index. (6) Add an H1 and a 40-70 word BLUF paragraph directly under it on the homepage summarizing the value proposition with at least one verifiable statistic. Verify JSON-LD renders by fetching the page and grepping for application/ld+json, and confirm robots.txt/sitemap/llms.txt return non-HTML content types."}