New AI Shopping Analytics — see if AI assistants put your products on the shortlist
Hartevo
Sign up

Make AI recommend your
company

Make AI recommend you Run it from your agent
The whole loop

Diagnose, build, distribute, prove. Other tools stop at a score; we write the fixes, publish them, and prove AI recommends you.

Western + Chinese AI

ChatGPT, Perplexity, Gemini, Grok, DeepSeek, plus China's Kimi, Doubao and GLM, in one report.

Inside your coding agent

Use 24 Skills and hosted MCP from Codex, Claude Code or Cursor.

ChatGPTcaptured today · real browser
“Best steam mop pads for a Shark? Top pick: MXZONE — washable and durable, fits Shark models. ✓”
MXZONE ✓

Ground-truth capture: the real answer, screenshotted.

Visibility trendyour share of AI answers, over weeks
Strong Building Weak

Now named by 6 of the engines we track

ChatGPT✓ Perplexity✓ Gemini✓ Grok✓ DeepSeek✓ Doubao✓

Illustrative dashboard

Search moved. Discovery didn't follow you.

Discovery collapsed from a page of blue links into a single AI answer. Either you're in it, or you're invisible.

Hartevo coordinating product evidence across Reddit, Medium, LinkedIn, X, Zhihu and Quora

Everyone ships now.

AI tools collapsed the cost of building. The hard part is no longer shipping an app; it's being found among everyone else who shipped one this week.

Buyers ask chatbots, not keyword boxes.

A growing share of product discovery happens inside ChatGPT, Gemini, and Grok answers, and in China, inside Doubao, Kimi, and GLM. If machines can't parse your site, you are simply not in the answer.

GEO is not SEO with a new name.

Rankings reward backlinks and dwell time. AI answers reward being fetchable, parseable, and citable: open robots rules, llms.txt, structured data, quotable copy.

More than a dashboard.

Most tools stop at a score. Hartevo diagnoses, builds and distributes, then proves it, in the surfaces buyers actually use.

ChatGPTChatGPT

Top pick: MXZONE — fits Shark✓

recommended in the AI answer
PerplexityPerplexity

Most-cited: MXZONE · 3 sources✓

recommended in the AI answer
GeminiGemini

For Shark mops, MXZONE✓

recommended in the AI answer
GrokGrok

Best value: MXZONE, fits Shark✓

recommended in the AI answer
DeepSeekDeepSeek

Recommended: MXZONE, washable✓

recommended in the AI answer
DoubaoDoubao

Best fit: MXZONE for Shark✓

recommended in the AI answer
Ground truth, not guesses

See the exact answer a buyer gets, captured in a real browser across ChatGPT, Perplexity, Google AI Overviews, Doubao and Kimi, by region.

Be the product AI recommends

Agentic-commerce ready: audit your catalog, export AI-shopping feeds, and track whether assistants put you on the shortlist.

Western + Chinese engines

One report covers ChatGPT, Perplexity, Gemini, Grok, DeepSeek, and China's Kimi, Doubao, GLM. The only self-serve tool that does both.

Distribute where AI looks

Auto-publish to the channels AI actually cites: Reddit, Medium, Dev.to, and China's Zhihu, Juejin, CSDN.

Run it from your agent

Drive the whole loop from Codex, Claude Code or Cursor through 24 Agent Skills and the hosted MCP.

Four moves, on a loop.

Diagnose, build, distribute, prove. Hand the complete loop to your coding agent through 24 Skills, hosted MCP or the CLI.

Agent commandMake AI recommend MXZONE
  1. Diagnose
  2. Build
  3. Distribute
  4. Prove
24 skills active
01

Diagnose

A 19-check read plus live citation tests across 8 Western and Chinese engines: a 0–100 score and your Share of Voice against competitors.

02

Build

Paste-ready llms.txt, JSON-LD and meta, authored from your own content, or opened as a pull request to your repo.

03

Distribute

Auto-publish to the channels AI actually cites: Reddit, Medium, Dev.to, and China's Zhihu, Juejin, CSDN.

04

Prove

Ground-truth capture in a real browser confirms what buyers actually see, across regions, week over week.

Make AI RecommendMXZONE
active campaign
  1. Baseline
  2. Diagnose
  3. Plan
  4. Execute
  5. Settle
  6. Verify
Mentioned0/7
Cited0/3
Recommended2/6
Recommendation trend0–100%

Published where AI actually looks.

We distribute to — and monitor — the open channels and source pools AI answers are drawn from. Western and Chinese, side by side.

What AI engines read

The same 19 signals AI crawlers effectively run when they decide whether your site is worth citing.

AI access32%

We confirm AI crawlers can reach you: robots.txt rules, llms.txt, and a live GPTBot fetch.

{ } ProductnameMXZONErating★ 4.8

Structure29%

The machine-readable scaffolding: JSON-LD, canonical, sitemap and the HTML basics.

best steam mop pads?Top pick: MXZONE ✓

Content26%

The text that gets quoted: title, heading depth, and real server-rendered words.

HartevoMake AI recommend youhartevo.com

Social15%

Rich shares everywhere: Open Graph, OG image, and Twitter cards.

24 Agent Skills + hosted MCP

Install the workflow library in Codex, Claude Code or Cursor, then run the whole loop where you already work.

CodexCodex
Claude CodeClaude Code
CursorCursor
Hartevo
Hosted MCP
"mcpServers": {
  "up-geo": {
    "type": "http",
    "url": "https://app.hartevo.com/mcp"
  }
}
CLI
$ geo scan yoursite.com
Agent Skills · 24
npx -y skills@latest add https://app.hartevo.com --skill "*" --agent codex --yes

Open Agent Skills & MCP docs →

Pricing without a sales call

A permanent free tier, and plans that scale on what you track: engines, prompts, brands. Never a percentage of your revenue.

What users say

Teams making themselves the answer AI recommends.

Michael Stone Michael StoneChief Executive Officer
“Hartevo helped us move from guessing our AI visibility to proving where our company appears, who outranks us, and what to fix next.”
Daniel Brooks Daniel BrooksVP of Growth
“The biggest result for us was clarity. Hartevo turned AI search visibility into a measurable company growth channel.”
Noah Bennett Noah BennettHead of Product
“For our app, Hartevo showed exactly why AI assistants were not recommending us and generated the technical fixes we needed to become citable.”
Min Jae Kim Min Jae Kim移动产品负责人
“我们真正补上的，是从“把 app 做出来”到“让 AI 读懂它、解释它、推荐它”之间的那段差距，这对增长影响很大。”
Kenji Tanaka Kenji Tanaka网站策略负责人
“网站在 AI 可读性上的问题被很快定位出来了，包括抓取、结构化数据、页面信息和内容缺口，后续优化方向一下子清楚了。”
Gregory Whitman Gregory WhitmanSenior Product Marketing Manager
“Hartevo showed whether AI engines could understand our product category, benefits, and competitors. That gave product marketing a much sharper roadmap.”
Ethan Miller Ethan MillerVP of Product Marketing
“The product impact was immediate: clearer positioning, better comparison content, and a stronger chance of being included in AI-generated shortlists.”
Hiroshi Sato Hiroshi Sato电商运营负责人
“做 AI 购物推荐这件事，我们以前基本没有抓手。现在至少能系统检查商品是否容易被引用、被理解，并进入 AI 推荐清单。”
Priya Rao Priya Rao客户洞察总监
“最有启发的一点，是它让我们真正站到买家的视角看品牌：AI 会怎么说我们、先推荐谁、又信任哪些来源。”
Sofia Romano Sofia RomanoCompetitive Intelligence Lead
“Hartevo connects multiple perspectives: technical readiness, content clarity, competitor share of voice, and buyer-facing AI recommendations.”
Mei Lin Mei Lin品牌总监
“以前我们只能凭感觉判断品牌在 AI 搜索里的表现，现在终于能清楚看到公司出现在哪里、被谁超过，以及下一步该修什么。”
Yuna Park Yuna Park市场总监
“对我们来说，最有价值的不是又多一个报表，而是拿到一套能直接执行的修复方案。低可见度这件事，终于有了清晰的改善路径。”
Arjun Mehta Arjun MehtaDirector of App Marketing
“We used Hartevo to improve how AI engines understand our app. The result was better positioning, cleaner citations, and a stronger presence across buyer prompts.”
Emma Clarke Emma ClarkeSEO Director
“Our website looked fine to humans, but Hartevo showed what AI crawlers could not read. The fix pack made the next steps obvious.”
Olivia Hart Olivia HartDigital Experience Manager
“The result was practical: better machine-readable pages, clearer AI summaries, and fewer blind spots in how our website is interpreted.”
Keiko Mori Keiko Mori产品策略总监
“这是我们第一次真正看清，产品在 AI 回答里到底是怎么被描述、比较和推荐的；看清之后，产品页和内容策略都跟着变了。”
Lily Dawson Lily DawsonEcommerce Director
“For our shop, Hartevo made AI shopping visibility measurable. We could see whether assistants included our products when buyers asked for recommendations.”
Ravi Santos Ravi SantosMarketplace Growth Manager
“Instead of optimizing only for search pages, we started optimizing for AI shopping answers. Hartevo gave us the evidence and the next actions.”
Margaret Ellis Margaret EllisDemand Generation Manager
“The value is not just the score. It shows the exact AI answer a customer may see, so our team can respond with real fixes.”
Anika Surya Anika Surya创始人兼 CEO
“现在这套流程终于闭环了：先发现问题，再生成修复方案、分发内容，最后验证 AI 引擎是不是真的开始推荐我们。”

FAQs

Questions, answered straight

What is GEO?

Generative Engine Optimization: making your site readable, parseable, and citable by AI systems (ChatGPT, Gemini, Grok, DeepSeek, and China's Doubao, Kimi, GLM) so they mention and recommend it. SEO targets ranking in a results page; GEO targets being the answer itself.

How is the score computed?

19 weighted checks in four pillars: AI access (32%), structure (29%), content (26%), social (15%). Each check is pass, warn, or fail. It measures machine readability, not how good your product is.

My app is a client-rendered SPA. Does that matter?

A lot. Most AI crawlers don't execute JavaScript. If your content only exists after JS runs, machines see an empty shell — typically a near-zero word count in our scan. We fix this with prerendering or static snapshots.

Do you cover the Chinese market?

Yes. We work in English and Chinese: Baidu basics, WeChat share cards, and launch posts on Chinese developer communities, alongside the global channels.

When do I see results?

Crawlers re-read technical fixes within days. Appearing inside AI answers usually follows within weeks, depending on your niche. Honest version: GEO compounds; it is not a switch.

Do you store my scan data?

No. We fetch your public pages, compute the report in memory, return it to your browser, and store nothing unless you opt into weekly monitoring.

Which AI engines do you cover?

Eight: ChatGPT, Perplexity, Gemini, Grok and DeepSeek, plus China's Kimi, Doubao and GLM. On Standard and up we also capture the consumer surfaces in a real browser: ChatGPT.com, Google AI Overviews, Copilot.

Do I need an account to start?

No. The scan, fix pack, promotion kit and one citation test are free with no login. Create a free account only when you want to save sites, track them weekly, and manage a subscription.

Make AI recommend you.

Run it from your agent, or hire ours.

Get started See the agent surfaces