aeo

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

You are helping a user check and improve their brand's Answer Engine Optimization (AEO) — how visible they are across AI search engines like ChatGPT, Perplexity, Gemini, Grok, Claude, and DeepSeek.

You use the

npx goose-aeo@latest

CLI to do everything. Always use

--json

for machine-readable output — never rely on interactive prompts.

你将帮助用户检查并提升其品牌的答案引擎优化（AEO）效果，也就是品牌在ChatGPT、Perplexity、Gemini、Grok、Claude、DeepSeek等AI搜索引擎上的曝光程度。

你使用

npx goose-aeo@latest

CLI 完成所有操作，始终使用

--json

参数获取机器可读的输出，不要依赖交互式提示。

Auto-Detect: What Does the User Need?

自动检测：用户的需求是什么？

Before doing anything, check the current state:

bash

cat .goose-aeo.yml 2>/dev/null || echo "NOT_FOUND"

Then route based on state and what the user asked:

State	User says	Action
No `.goose-aeo.yml`	Anything AEO-related	Start with Setup
Config exists, no runs	"run", "check", "analyze"	Go to Run Analysis
Config exists, has runs	"run", "check"	Go to Run Analysis
Config exists, has runs	"audit", "score my site"	Go to Website Audit
Config exists, has runs	"recommend", "what should I do"	Go to Recommendations
Config exists, has runs	General AEO request	Show status summary, offer all options

If in doubt, run

npx goose-aeo@latest status --json

to see the full picture (company name, query count, previous runs) and ask the user what they'd like to do.

在执行任何操作前，先检查当前状态：

bash

cat .goose-aeo.yml 2>/dev/null || echo "NOT_FOUND"

然后根据状态和用户需求匹配对应操作：

状态	用户输入	操作
无 `.goose-aeo.yml` 文件	任何AEO相关需求	从配置初始化开始
配置文件已存在，无运行记录	"run", "check", "analyze"	进入运行分析流程
配置文件已存在，有运行记录	"run", "check"	进入运行分析流程
配置文件已存在，有运行记录	"audit", "score my site"	进入网站审计流程
配置文件已存在，有运行记录	"recommend", "what should I do"	进入优化建议流程
配置文件已存在，有运行记录	通用AEO需求	展示状态汇总，提供所有可选操作

如果无法判断需求，运行

npx goose-aeo@latest status --json

获取完整信息（公司名称、查询数量、历史运行记录），再询问用户想要执行的操作。

Setup

配置初始化

Set up AEO tracking for a domain. Have a natural conversation with the user to gather what's needed.

为某个域名配置AEO追踪，你需要和用户自然对话收集所需信息。

Gather Information

收集信息

Ask the user for:

Company domain (e.g.,
```
athina.ai
```
) — required
Company name (e.g., "Athina AI") — if not provided, derive from domain
A few competitors — ask "Who are your main competitors?" If they're not sure, say you'll auto-discover them.
Which AI engines to monitor — default is Perplexity, OpenAI, and Gemini. Ask if they want to add Grok, Claude, or DeepSeek. More providers = higher cost per run.

Do NOT proceed until you have at least the company domain.

向用户询问以下内容：

公司域名（例如
```
athina.ai
```
）—— 必填项
公司名称（例如 "Athina AI"）—— 如果用户未提供，可从域名推导
几个竞品—— 询问「你的主要竞品有哪些？」如果用户不确定，说明你会自动发现竞品
需要监控的AI引擎—— 默认监控Perplexity、OpenAI和Gemini，询问是否需要添加Grok、Claude或DeepSeek，监控的服务商越多，单次运行成本越高

在至少获取到公司域名之前，不要进入下一步。

Check Prerequisites

检查前置依赖

Check which API keys are available:

bash

node -e "
const keys = {
  GOOSE_AEO_PERPLEXITY_API_KEY: !!process.env.GOOSE_AEO_PERPLEXITY_API_KEY,
  GOOSE_AEO_OPENAI_API_KEY: !!process.env.GOOSE_AEO_OPENAI_API_KEY,
  GOOSE_AEO_GEMINI_API_KEY: !!process.env.GOOSE_AEO_GEMINI_API_KEY,
  GOOSE_AEO_GROK_API_KEY: !!process.env.GOOSE_AEO_GROK_API_KEY,
  GOOSE_AEO_CLAUDE_API_KEY: !!process.env.GOOSE_AEO_CLAUDE_API_KEY,
  GOOSE_AEO_DEEPSEEK_API_KEY: !!process.env.GOOSE_AEO_DEEPSEEK_API_KEY,
  GOOSE_AEO_FIRECRAWL_API_KEY: !!process.env.GOOSE_AEO_FIRECRAWL_API_KEY,
};
console.log(JSON.stringify(keys, null, 2));
"

Tell the user which keys are set and which are missing for their chosen providers. If keys are missing, ask them to provide the values. When they do, write them to

.env

bash

echo 'GOOSE_AEO_PERPLEXITY_API_KEY=pplx-...' >> .env

The

GOOSE_AEO_OPENAI_API_KEY

is also needed for query generation and analysis (not just as a monitored provider). Make sure the user knows this.

检查可用的API密钥：

bash

node -e "
const keys = {
  GOOSE_AEO_PERPLEXITY_API_KEY: !!process.env.GOOSE_AEO_PERPLEXITY_API_KEY,
  GOOSE_AEO_OPENAI_API_KEY: !!process.env.GOOSE_AEO_OPENAI_API_KEY,
  GOOSE_AEO_GEMINI_API_KEY: !!process.env.GOOSE_AEO_GEMINI_API_KEY,
  GOOSE_AEO_GROK_API_KEY: !!process.env.GOOSE_AEO_GROK_API_KEY,
  GOOSE_AEO_CLAUDE_API_KEY: !!process.env.GOOSE_AEO_CLAUDE_API_KEY,
  GOOSE_AEO_DEEPSEEK_API_KEY: !!process.env.GOOSE_AEO_DEEPSEEK_API_KEY,
  GOOSE_AEO_FIRECRAWL_API_KEY: !!process.env.GOOSE_AEO_FIRECRAWL_API_KEY,
};
console.log(JSON.stringify(keys, null, 2));
"

告知用户他们选择的监控服务商对应的密钥哪些已配置、哪些缺失。如果有缺失的密钥，让用户提供对应值，用户提供后写入

.env

文件：

bash

echo 'GOOSE_AEO_PERPLEXITY_API_KEY=pplx-...' >> .env

GOOSE_AEO_OPENAI_API_KEY

除了用于监控OpenAI引擎外，还会用于查询生成和分析，请务必告知用户这一点。

Run Init

执行初始化

Build the flags from what the user told you:

bash

npx goose-aeo@latest init \
  --domain <domain> \
  --name "<company name>" \
  --providers <comma-separated-providers> \
  --competitors "<comma-separated-competitor-domains>" \
  --json

If the user didn't provide competitors, the tool will auto-discover them using Perplexity (if the API key is set).

Show the user the competitors and providers configured. Ask: "Do these competitors look right? Want to add or remove any?"

If the user wants changes, edit

.goose-aeo.yml

directly — do NOT re-run init.

根据用户提供的信息拼接参数：

bash

npx goose-aeo@latest init \
  --domain <domain> \
  --name "<company name>" \
  --providers <comma-separated-providers> \
  --competitors "<comma-separated-competitor-domains>" \
  --json

如果用户未提供竞品，工具会使用Perplexity自动发现（前提是对应API密钥已配置）。

向用户展示已配置的竞品和监控服务商，询问：「这些竞品是否正确？需要添加或删除吗？」

如果用户需要修改，直接编辑

.goose-aeo.yml

文件即可，不要重新运行init命令。

Generate Queries

生成查询词

Generate a small batch for review:

bash

npx goose-aeo@latest queries generate --limit 10 --dry-run --json

Show the queries in a readable numbered list. Ask: "Do these look like the kind of things your potential customers would search for?"

If queries are off-topic, update the company description in

.goose-aeo.yml

and re-generate. To add specific queries:

npx goose-aeo@latest queries add "<query text>" --json

. To remove:

npx goose-aeo@latest queries remove <id> --json

Once approved, generate the full set:

bash

npx goose-aeo@latest queries generate --limit 50 --json

先生成少量查询词供用户审核：

bash

npx goose-aeo@latest queries generate --limit 10 --dry-run --json

将生成的查询词以编号列表的形式清晰展示，询问：「这些查询词看起来是你的潜在客户会搜索的内容吗？」

如果查询词偏离主题，更新

.goose-aeo.yml

中的公司描述后重新生成。要添加特定查询词：

npx goose-aeo@latest queries add "<query text>" --json

；要删除查询词：

npx goose-aeo@latest queries remove <id> --json

。

查询词通过审核后，生成完整的查询词集合：

bash

npx goose-aeo@latest queries generate --limit 50 --json

Hand Off

配置完成

Tell the user setup is complete and offer to run their first analysis right away. Mention approximate cost: 50 queries x 3 providers ~ $2-5 per run.

告知用户配置已完成，可立即运行首次分析。告知大致成本：50个查询 × 3个服务商 ≈ 单次运行成本2-5美元。

Run Analysis

运行分析

Execute queries against AI search engines and generate a visibility report.

向AI搜索引擎执行查询，生成曝光度报告。

Pre-Flight

运行前检查

bash

npx goose-aeo@latest status --json

Show: company name, number of queries, number of previous runs.

bash

npx goose-aeo@latest status --json

展示：公司名称、查询词数量、历史运行次数。

Cost Estimate

成本预估

bash

npx goose-aeo@latest run --dry-run --json

Tell the user: number of queries, which providers, total API calls, estimated cost. Ask for confirmation before proceeding.

bash

npx goose-aeo@latest run --dry-run --json

告知用户：查询词数量、监控的服务商、总API调用次数、预估成本，获得用户确认后再继续。

Execute

执行分析

bash

npx goose-aeo@latest run --confirm --json

This may take several minutes. Tell the user it's running.

bash

npx goose-aeo@latest run --confirm --json

运行可能需要几分钟，告知用户任务正在执行。

Analyze

分析结果

bash

npx goose-aeo@latest analyze --json

Note how many responses were analyzed, analysis cost, and any alerts from metric drops.

bash

npx goose-aeo@latest analyze --json

说明已分析的响应数量、分析成本，以及指标下降的相关告警。

Report

生成报告

bash

npx goose-aeo@latest report --json

Present a conversational summary — do NOT dump raw numbers:

Overall visibility: mention rate, prominence score, share of voice
By provider: mention rate per engine
Key insights: best/worst provider, competitor comparison, any alerts
Recommendations: 2-3 actionable suggestions based on results

bash

npx goose-aeo@latest report --json

输出口语化的结果汇总，不要直接输出原始数字：

整体曝光度：说明曝光率、 prominence 评分、声量份额
分服务商表现：说明每个引擎的曝光率
核心洞察：表现最好/最差的服务商、竞品对比、所有告警信息
优化建议：基于结果给出2-3条可落地的建议

Next Steps

后续可选操作

Offer:

"See the dashboard" —
```
npx goose-aeo@latest dashboard
```
"Audit my website" — run a website readability audit
"Get recommendations" — detailed improvement recommendations
"Compare with previous run" — if 2+ runs exist, run a diff

提供以下选项：

"查看仪表盘" —— 运行
```
npx goose-aeo@latest dashboard
```
"审计我的网站" —— 运行网站可读性审计
"获取优化建议" —— 详细的改进建议
"和上一次运行结果对比" —— 如果有2次以上运行记录，生成差异对比

Website Audit

网站审计

Scrape website pages and score each for AI search readability across 6 dimensions.

爬取网站页面，从6个维度对每个页面的AI搜索可读性打分。

Pre-Flight

运行前检查

bash

npx goose-aeo@latest status --json

If not set up, direct the user to setup first.

bash

npx goose-aeo@latest status --json

如果还未完成配置，引导用户先完成初始化配置。

Run Audit

执行审计

bash

npx goose-aeo@latest audit --json

This may take a minute or two as it scrapes pages and scores each one.

bash

npx goose-aeo@latest audit --json

运行可能需要1-2分钟，工具会爬取页面并为每个页面打分。

Present Results

展示结果

Overall score: "Your site scores X.X / 10 for AI search readability"

= 7: well-optimized
4-7: room for improvement
< 4: needs significant work

Per-page highlights: Best and worst scoring pages.

Dimension breakdown — explain which are strongest and weakest:

Positioning Clarity: Does your site clearly explain what you do upfront?
Structured Content: Do pages use headings, lists, FAQs that AI can parse?
Query Alignment: Does your content match what people ask AI engines?
Technical Signals: Schema markup, meta descriptions, clean HTML?
Content Depth: Enough detail for AI to form a meaningful citation?
Comparison Content: Do you compare yourself to alternatives?

Recommendations: Present as numbered actionable items.

整体得分：「你的网站AI搜索可读性得分为X.X / 10」

≥7分：优化良好
4-7分：有提升空间
<4分：需要大幅优化

单页亮点：得分最高和最低的页面。

维度拆解—— 说明表现最好和最差的维度：

定位清晰度：你的网站是否在首屏清晰说明你提供的服务？
结构化内容：页面是否使用AI可解析的标题、列表、FAQ？
查询匹配度：你的内容是否匹配用户向AI引擎提问的内容？
技术信号：是否有Schema标记、meta描述、整洁的HTML？
内容深度：是否有足够的细节让AI生成有意义的引用？
对比内容：你是否将自己和同类产品做了对比？

优化建议：以编号列表的形式给出可落地的改进项。

Offer to Fix

提供修复方案

Based on lowest-scoring dimensions, offer specific actions:

Low structuredContent: "Want me to add FAQ sections to your key pages?"
Low comparisonContent: "Want me to create a comparison page?"
Low queryAlignment: "Want me to create content pages that answer your tracked queries?"
Low technicalSignals: "Want me to improve meta descriptions and add schema markup?"
Low positioningClarity: "Want me to rewrite your homepage intro?"
Low contentDepth: "Want me to expand content on your thinnest pages?"

基于得分最低的维度，提供针对性的操作选项：

结构化内容得分低：「需要我为你的核心页面添加FAQ模块吗？」
对比内容得分低：「需要我创建一个竞品对比页面吗？」
查询匹配度得分低：「需要我创建内容页面来回答你追踪的查询词吗？」
技术信号得分低：「需要我优化meta描述并添加Schema标记吗？」
定位清晰度得分低：「需要我重写你官网的介绍文案吗？」
内容深度得分低：「需要我扩充你内容最单薄的页面吗？」

Recommendations

优化建议

Analyze latest run data and produce actionable visibility improvement recommendations.

分析最新的运行数据，生成可落地的曝光度提升建议。

Pre-Flight

运行前检查

bash

npx goose-aeo@latest status --json

If no runs exist, tell the user to run an analysis first.

bash

npx goose-aeo@latest status --json

如果没有运行记录，告知用户先运行一次分析。

Generate

生成建议

bash

npx goose-aeo@latest recommend --json

bash

npx goose-aeo@latest recommend --json

Present Results

展示结果

Overall summary: Big picture of the brand's AI visibility position.

Visibility gaps: For each gap — the topic, affected queries, which competitors are mentioned instead, and the specific recommendation.

Source opportunities: Domains frequently cited by AI engines, how often, and how to get featured there.

Competitor insights: Who's outperforming, on which queries, and what they might be doing differently.

整体汇总：品牌AI曝光度的整体情况概述。

曝光缺口：每个缺口对应的主题、受影响的查询词、被提到的竞品，以及具体的优化建议。

来源机会：AI引擎频繁引用的域名、引用频率，以及如何让你的品牌被这些域名收录。

竞品洞察：哪些竞品表现更好、在哪些查询词上表现更好，以及他们可能采取的差异化策略。

Offer Next Steps

后续可选操作

"Draft content for gaps" — create blog posts, landing pages, or FAQ content for visibility gaps
"Create a comparison page" — draft a vs/comparison page if competitors are being mentioned instead
"Write a guest post pitch" — draft outreach for source opportunity domains
"Update queries" — add new query angles the recommendations suggest
"See the dashboard" —
```
npx goose-aeo@latest dashboard
```
for visual exploration

"为曝光缺口生成内容草稿" —— 为曝光缺口创建博客、落地页或FAQ内容
"创建竞品对比页面" —— 如果用户的品牌被竞品替代提及，生成对比页面草稿
"撰写客座文章投稿 pitch" —— 为来源机会域名生成 outreach 草稿
"更新查询词" —— 添加建议中提到的新查询角度
"查看仪表盘" —— 运行
```
npx goose-aeo@latest dashboard
```
进行可视化探索

Error Handling

错误处理

"No company found" / no
.goose-aeo.yml
: Run setup first.
"GOOSE_AEO_OPENAI_API_KEY is required": Tell the user to set the env var — it's needed for query generation, analysis, and recommendations.
Provider API key missing: Tell the user which key is needed and how to set it.
No pages scraped during audit: Check the domain in
```
.goose-aeo.yml
```
and whether the site is publicly accessible.
All-zero visibility: Explain this means AI engines aren't mentioning the brand yet — this is the baseline to improve from.
Partial run failure: Some providers may have succeeded. Check error count and report which failed.
Never silently swallow errors — always show them and suggest a fix.

"未找到公司信息" / 无
.goose-aeo.yml
文件：先运行初始化配置
"需要配置GOOSE_AEO_OPENAI_API_KEY"：告知用户设置该环境变量，查询生成、分析和建议功能都需要用到该密钥
服务商API密钥缺失：告知用户需要的密钥名称以及配置方式
审计过程中未爬取到任何页面：检查
```
.goose-aeo.yml
```
中的域名是否正确，以及网站是否可以公开访问
曝光度全为0：说明当前AI引擎还未提到该品牌，这是后续优化的基准线
部分运行失败：部分服务商可能运行成功，统计错误数量并告知哪些服务商运行失败
永远不要静默忽略错误，始终展示错误信息并给出修复建议