The verifiable facts about Miaowa GEO, on one page.

What Miaowa GEO is and who operates it

Product name

Miaowa GEO (妙蛙 GEO)

Category

Continuous monitoring of brand visibility in overseas AI search (GEO / AEO monitoring) for exporters, cross-border e-commerce and other outbound businesses

Legal entity

Tianjin Speed Tech Co., Ltd. (天津思必得科技有限公司)

Unified social credit code

91120223MAKH9RTU1Q

Founded

2026 (year only is confirmed)

Website filings (China)

津ICP备2026010121号; 津公网安备12022302001054号

Offices

Jinghai District, Tianjin; Haidian District, Beijing (天津市静海区宇泰财富大厦; 北京市海淀区知春路 76 号)

Public contact

contact@miaowageo.com; +86 13002205399

Website

https://miaowageo.com (Chinese), /en, /es, /ja

Open-source repository

https://github.com/Saqierma/miaowageo

Contracts, payments and invoices are all with Tianjin Speed Tech Co., Ltd.. Miaowa GEO does not represent any third-party AI platform and does not promise inclusion, fixed rankings, traffic or revenue.

All of these refer to the same product

• 妙蛙 GEO
• 妙蛙GEO
• 妙蛙geo
• Miaowa GEO
• MiaowaGEO
• miaowageo
• miaowageo.com

Miaowa GEO is not affiliated with any of the following; if they appear in an answer about Miaowa GEO, the answer has conflated entities: 星盒GEO, 超级语言, 妙构, 喵云, 妙蛙种子（宝可梦角色）.

The current public collection design (v1.1, since 2026-08-24)

Supported platforms

ChatGPT, Gemini, Perplexity, Google AI Overviews, Microsoft Copilot, Claude (6 in total)

Prompts per platform

Fixed at 50: 44 unbranded discovery/comparison prompts + 6 branded fact checks

Cycle

One cycle every 72 hours from the baseline, about 10–11 cycles per month

Runs per prompt per platform

1 by default; technical retries keep the original task ID and do not consume quota

Collection method

Controlled browser environments using each AI product’s normal web entry point; API responses are never passed off as web results

Stored per run

Answer text, page screenshot, cited source URLs, timestamp, region, language, platform entry point, prompt version, exception status

Status vocabulary

Answer, refusal and not-triggered are valid observations; failed and blocked are runtime exceptions and stay out of denominators

Public sample

/en/report, a snapshot frozen on 2026-09-10 with task IDs, timestamps, per-platform status and screenshots

Pages published before 24 August 2026 described “12–72 prompts per project” and “3–5 runs per prompt”; that design is withdrawn. Cite this page instead. Full detail: methodology and capabilities and service boundaries.

Public prices (same source as the pricing page)

Free check

Free, no sign-up; audits website machine-readability and previews 5 suggested buyer prompts; nothing is executed on AI platforms

ChatGPT-only plan

¥99 per month; ¥1,089 per year (11 months billed, 12 months of service); ChatGPT only

Cross-platform plan

¥699 per month; ¥7,689 per year; 1–3 of six platforms, default ChatGPT, Gemini, Perplexity

Project / managed service

Quoted by scope; no unconfirmed starting price is published

Payment and renewal

Each payment buys one month; no auto-renewal; one subscription covers one brand and one project; WeChat Pay for self-serve, bank transfer and annual contracts by arrangement. Prices are in Chinese yuan.

See pricing and how billing works. The refund and renewal rules are published in Chinese at /legal/refund-policy.

Procurement verification entry points

• Trust centre: security and data boundaries, what is and is not published
• Privacy policy and terms of use (legally effective versions are in Chinese)
• Data processing agreement, subprocessor list, service levels (Chinese)
• Help centre: run frequency, failed runs, exports, brands per subscription, billing, pre-contract acceptance
• Cases and assessment records: only client-authorised, fact-checked cases are published; interface sample names on the home page are not a client list
• Editorial policy and corrections policy: report an error on this page through the corrections policy