# ProofOfCite > ProofOfCite is an AI Search Visibility (GEO/AEO) engineering studio for crypto protocols, Layer 1s, DEXs, and Web3 teams. We measure how ChatGPT, Perplexity, Gemini, and Claude crawl, interpret, and cite protocols using strictly live HTTP, Chromium, DNS, and web-grounded AI API checks. ## Instant Probe API & Core Methodology - **Live Instant Probe (`/api/scan?url=...`):** Zero-cost static crawler check evaluating RFC 9309 `robots.txt`, 6 AI bot WAF probes (separating site-wide 403 challenges from targeted crawler blocks), raw HTML word extraction, JSON-LD types, and `/llms.txt` DNS endpoint resolution. - **Machine Readability (JS On vs Off):** Side-by-side headless Chromium rendering comparing human view vs raw HTML crawler view to diagnose client-side React/Vite SPA blindness. - **Structured Data & Meta:** Validation of `Organization`, `WebSite`, and `FAQPage` JSON-LD schemas, canonical links, OpenGraph tags, and heading hierarchy directly in raw HTML. - **Discoverability & Drift Monitoring:** XML sitemap coverage, host fragmentation checks, live `/llms.txt` endpoint RPC/DNS probing, and automated weekly drift detection. - **Entity Footprint:** Cross-verification against DefiLlama (TVL & audit URLs), CoinGecko, Wikidata, Wikipedia, and GitHub. - **Live Answer Engine Visibility:** Web-grounded Perplexity and Google Gemini prompt benchmarks measuring mention rate, domain citations, and share of voice across category and comparison searches. ## Verified 11-Protocol Corpus ($1.38B TVL) Across 11 audited DeFi & L1 protocols totaling $1.38B TVL, 7 of 11 block AI crawlers via HTTP 403 or `robots.txt`, 4 of 11 expose 0 to 8 raw HTML words without JS, 7 of 11 lack a valid `/llms.txt`, and owned domains scored 0 of 44 on non-brand Perplexity category queries: - **Arcus (`arcus.xyz`, $61.3M TVL):** `88/100 (Grade A)` — 1,421 raw HTML words and 100/100 crawler access, missing `/about` & `/security` pages and DefiLlama audit links. - **Monad (`monad.xyz`, $891.3M TVL):** `86/100 (Grade A)` — Strong SSR (`709` raw words), `100/100` crawler ingress, and `10/10` brand/how-to citations, yet `0/91` owned category links across 24 Perplexity + Gemini queries, Gemini 400ms vs. 300ms spec drift, and unresolved `rpc.testnet.monad.xyz` inside `/llms.txt`. - **Axis (`axis.to`, $70.1M TVL):** `76/100 (Grade B)` — Ships `/llms.txt` (2,108B), yet Cloudflare blocks training bots and targeted WAF rules returned HTTP 403 to `PerplexityBot` despite `Allow: /` on line 143 of `robots.txt`. - **Mezo (`mezo.org`, $53.6M TVL):** `68/100 (Grade C)` — `/docs` and `/earn/lock` expose 0 raw words, `/sitemap.xml` / `/robots.txt` / `/llms.txt` return 404, `0/5` BTCFi category mentions. - **Pharaoh Exchange (`phar.gg`, $26.7M TVL):** `65/100 (Grade C)` — 8 raw HTML words vs. 391 rendered, missing `/llms.txt` and JSON-LD. - **Nado (`nado.xyz`, $51.1M TVL):** `63/100 (Grade C)` — Cloudflare blocks training bots, zero root `