aeorank scan

Scan a website URL and generate an AEO score with a dimension breakdown and the 8 files the published CLI (aeorank-cli@0.1.1) actually writes.

Usage

aeorank scan <url> [options]

Arguments

Argument	Description
<url>	The website URL to scan (required)

Options

Flag	Default	Description
--format <type>	human	Output format: human (colored terminal) or json (machine-readable)
--output <dir>	./aeorank-output	Directory to write generated files
--config <path>	—	Path to configuration file
--max-pages <n>	200	Maximum pages to crawl
--no-files	off	Skip writing generated files
--overwrite	off	Replace existing output files
--browser / -b	off	Use Playwright for JavaScript-rendered pages
--pillar <name>	—	Filter dimensions to one pillar
--page <path>	—	Show score for a specific page path
--fail-on-crawler-block	off	Fail the PR if GPTBot is blocked. Exit 2 if GPTBot, ClaudeBot, PerplexityBot, or Google-Extended is disallowed in robots.txt. A missing robots.txt is unknown, not blocked.

Examples

Basic scan

npx aeorank-cli scan https://example.com

Fail CI when an AI crawler is blocked

npx aeorank-cli scan https://example.com --fail-on-crawler-block --format json --no-files

Exit codes:

Code	Meaning
0	Scan succeeded (unknown crawlers do not fail)
1	Scan or I/O error
2	--fail-on-crawler-block and a gated bot is disallowed

JSON output for CI

npx aeorank-cli scan https://example.com --format json --no-files

The JSON object is the contract the GitHub Action consumes next. Always present:

• version — CLI version from package.json (one source of truth)
• score / grade
• dimensionCount — dimensions.length (currently 36)
• generatedFiles — names of files actually written (currently 8)
• crawlerAccess — per-bot map: allow | block | unknown
• crawlerGate — { checkedBots, blockedBots, unknownBots, failed, robotsTxt }
  • failed is true only when a checked bot is disallowed
  • robotsTxt: "missing" means 404 / no file — status unknown, not blocked

npx aeorank-cli scan https://example.com --format json --no-files | jq '.crawlerAccess.GPTBot'

Output

Human format (default)

1. AI crawler table — GPTBot, ClaudeBot, PerplexityBot, Google-Extended → allow / block / unknown
2. Overall AEO score (0–100) and letter grade, plus the true dimension and file counts
3. 36-dimension breakdown grouped by pillar
4. Actionable next-step recommendations

JSON format

See fields above. crawlerAccess uses allow / block / unknown (not allowed / disallowed) so Actions can consume it without remapping.

Honest claims

Crawler allowlists plus a CI gate prevent GPTBot (and peers) from staying blocked. llms.txt is agent-docs hygiene, not a citation guarantee.