# WebsiteIQ

> WebsiteIQ is an evidence-first website auditor for technical SEO, structured data, accessibility, performance, trust signals, crawlability, and machine discovery. It turns publicly observable behavior into a bounded, reproducible report for owners, agencies, and software agents. Core scoring is deterministic and LLM-free; heuristics are versioned and labeled.

Operator: Osborn Ventures Inc.
Contact: https://websiteiq.xyz/contact/
Canonical: https://websiteiq.xyz/

## Human documentation

- [Website auditor checks](https://websiteiq.xyz/website-audits/): scope, modules, limits, and report delivery.
- [Public service statistics](https://websiteiq.xyz/status/): Worker-backed aggregate audit totals, current public limits, and API availability.
- [Methodology 1.2](https://websiteiq.xyz/methodology/): observations, derivations, heuristics, and uncertainty.
- [WebsiteIQBot policy](https://websiteiq.xyz/crawler/): identity and robots-first crawl contract.
- [White-label reports](https://websiteiq.xyz/white-label-reports/): agency delivery.
- [x402 status](https://websiteiq.xyz/x402/): live API/MCP-only x402 v2 USDC settlement on Base.
- [Real audit report](https://websiteiq.xyz/sample-report/): published 10-page WebsiteIQ audit of websiteiq.xyz with private delivery identifiers removed.
- [Privacy notice](https://websiteiq.xyz/privacy/): collection, use, no-sale statement, retention, and rights.

## Machine interfaces

- OpenAPI description: https://websiteiq.xyz/api/openapi.json
- RFC 9727 API catalog: https://websiteiq.xyz/.well-known/api-catalog
- Agent access and x402 credential guidance: https://websiteiq.xyz/auth.md
- MCP Server Card: https://websiteiq.xyz/.well-known/mcp/server-card.json
- Agent Skills Discovery v0.2 index: https://websiteiq.xyz/.well-known/agent-skills/index.json
- Website audit skill: https://websiteiq.xyz/.well-known/agent-skills/website-audit/SKILL.md
- Official MCP Registry record (`io.github.OsbornVentures/websiteiq`, active version 1.4.0): https://registry.modelcontextprotocol.io/v0.1/servers?search=io.github.OsbornVentures%2Fwebsiteiq
- RFC 8288 discovery: the homepage response advertises the API catalog with the registered `api-catalog` Link relation.
- Public audit-check catalog: https://websiteiq.xyz/api/checks/catalog
- Public aggregate statistics: https://websiteiq.xyz/api/stats
- OKF v0.2 entry: https://websiteiq.xyz/okf/
- OKF manifest: https://websiteiq.xyz/okf/manifest.json

The guarded MCP endpoint is advertised through the API catalog and OKF agent instructions rather than placed in the crawl corpus.

## Important limits

- WebsiteIQBot obeys robots.txt, stays on the submitted host, rejects private targets, and stops on HTTP 402.
- Free browser, API, and MCP access: one comprehensive single-page audit per requester IP and per target website each day; no account, email, or contact consent is required.
- Free-audit abuse controls cross-reference pseudonymous requester-IP, target-website, and requester-target-pair counters. Human limit responses give a concise IP/site reason; API and MCP errors also include that reason plus a robots-aware, sitemap-priced payable x402 challenge.
- Paid whole-site execution is available through API and MCP x402: standard audits cost $0.01 through 10 pages plus $0.001 per page above 10; white-label audits cost $0.02 through 10 pages plus $0.002 per page above 10. The normal website has no paid checkout path.
- `POST /api/estimate` and MCP `estimate_audit` read robots.txt first, count a bounded robots-permitted sitemap tree, and return standard and white-label quotes before audit crawling. Estimation itself does not collect payment or start an audit; `POST /api/paid-audits` or the quote-specific paid resource issues the live x402 challenge and starts work only after settlement.
- White-label quote identity requires agency or business name, presenter name, phone, email, and consent to process those agency contact fields. Optional CTA and marketing inputs replace WebsiteIQ's report contact button only after a premium white-label payment settles; WebsiteIQ attribution, methodology, evidence, and uncertainty labels remain visible.
- Real reports use unlisted token-bearing links, are noindex/no-store, are retained until deletion is requested, and are deliberately absent from public discovery.
- E-E-A-T and YMYL outputs are WebsiteIQ heuristics, not Google scores or ranking predictions.
- x402 v2 USDC settlement on Base and queued batch execution are live for the documented paid API and MCP resources. An abuse-limited agent receives the exact reason plus a robots-aware, sitemap-priced payable challenge.