Six free field instruments

Grade the web
agents see.

One URL. Six separate verdicts. No blended score that hides what failed.

Choose the question.

Discovery, readability, actionability, and trust.

Crawling, indexing, and page metadata.

Answer structure, decision facts, and source signals.

Runtime discovery and tool quality.

A cross-agent access matrix.

Locale discovery and parity.

Public evidence only. Every finding names what the scanner fetched or observed.

Separate denominators. SEO, access, WebMCP, and locale readiness are different jobs.

No tool execution. The WebMCP grader discovers and inspects tools without calling them.