Get Started
DocsTrack 5aeo
/llms.txt
AEO / llms.txt4 min readbrevio-core v0.11.1

AEO Scorecard (brevio-core aeo), /llms.txt & Answer-Engine Readiness

Plain-English Summary (For Anyone)

Check whether AI answer engines can find, read and cite your pages

Answer Engine Optimization (AEO) is about making pages easy for AI search to find and quote. `brevio-core aeo` reads a folder of built HTML pages, with its robots.txt and llms.txt, and prints a score out of 100 for each page, worst first, with the gaps under it.

Why this matters:
It reads local files only: it fetches nothing, and given a URL it stops and tells you to save the page first.
It tells you when robots.txt shuts out AI crawlers, and which site files are missing.
It can fail a build with `--min SCORE`, so a regression in your site files is caught in CI.
When to use:Run it on your static build folder, locally or in CI, with `brevio-core aeo out/ --min 75`.
Technical Contract (For Engineers)
Line-level rules
CLI Invocation Signature:
brevio-core aeo <folder-or-page> [--min SCORE] [--format text|markdown|json] [--generate-llms]
Latency / Execution
Not measured for 0.11.1
Token Cost
No model call. No network: it reads local files only.
Exit Codes
0 = every page scores --min or more (default 0: report only); 1 = a page scores below --min; 2 = usage error (a URL, a missing path, no .html page, a bad --min)
Artifacts & Files Generated:
llms.txt (with --generate-llms)robots.txtskill.md.well-known/agent-card.json
Visual Execution Pipeline
5-Stage Progressive Flow
01Local files
Read
Every .html page under the folder, plus robots.txt, llms.txt, skill.md and the agent card
02Part 1
Discovery
robots.txt and AI crawlers, llms.txt, a canonical URL
03Part 2
Structure
Heading tree, a table, JSON-LD, enough text to quote
04Part 3
Token economics
How much of the page is markup, and whether a Markdown copy exists
05Part 4
Capability signals
skill.md, an agent card or an MCP descriptor

Why Answer Engine Optimization Matters

In Plain English:More people ask an AI for an answer than click ten links. If an answer engine cannot read your page, it cannot quote it.
AEO is about whether the pages of a site can be found, read and cited by answer engines. A robots.txt that shuts out AI crawlers, a page that needs scripts to show its text, or no structured data are the kind of gaps `brevio-core aeo` lists. It is a separate command from the pull-request gate.

The Scorecard: Four Parts, Scaled to 100

In Plain English:Each page gets a score out of 100, and under it the gaps, the first ones first.
`brevio-core aeo` scores four parts and scales the result to a score out of 100: 1. **Discovery**: whether `robots.txt` shuts out the AI crawlers, `llms.txt` and its per-page token counts, and a canonical URL. 2. **Structure**: text a crawler sees without running scripts, a heading tree, a table, JSON-LD structured data and enough text to quote. 3. **Token economics**: how much of the page is markup, the page's size and whether a Markdown copy exists. 4. **Capability signals**: `skill.md`, an agent card (`.well-known/agent-card.json`) or an MCP descriptor. Its text and Markdown reports start with what the site files say, and `--format json` adds each part's score. This site's own /scan page has its own scoring and is not what `brevio-core aeo` prints.

`llms.txt` and `--generate-llms`

In Plain English:A plain-text index of your pages, written for AI models to read in one request.
`brevio-core aeo` checks that the site root has an `llms.txt`, and reports its per-page token counts. `--generate-llms` writes a missing `llms.txt` at the target folder's root from its HTML pages before scoring. A Markdown copy beside a page (`page.md` or `page.html.md`) counts as the page's Markdown mirror.

Running It in CI (`--min`)

In Plain English:Fail the build if someone accidentally blocks AI crawlers or removes your site files.
Run it after the build step, on the build output: `pip install brevio-core==0.11.1` then `brevio-core aeo out/ --min 75 --format markdown >> "$GITHUB_STEP_SUMMARY"`. If any page scores below the minimum the step fails with exit code 1. The default `--min` is 0, which only reports.
AEO Check in a Workflow
Fixed (After)Exit 0
- run: npm run build
- run: pip install brevio-core==0.11.1
- run: brevio-core aeo out/ --min 75 --format markdown >> "$GITHUB_STEP_SUMMARY"

Interactive Rule Explorer

Explore the brevio-core rules and load their examples into the panel on the right.

129 / 129 rules