Skip to content
Kitbay
Skill

ai-visibility

MITUpdated on Nov 9, 2026606GitHub starsIncludes scripts

A skill from geo-doctor, a Chinese and English toolkit for getting products described correctly by AI assistants, that measures whether assistants such as ChatGPT, DeepSeek, Doubao, Qwen, Kimi, Perplexity and Google AI Overviews mention, recommend or cite a brand.

What the skill teaches

A skill from geo-doctor, a Chinese and English toolkit for getting products described correctly by AI assistants, that measures whether assistants such as ChatGPT, DeepSeek, Doubao, Qwen, Kimi, Perplexity and Google AI Overviews mention, recommend or cite a brand. The agent builds a set of questions real users would ask, split by type, then records answers sampled by hand in each assistant's app, with each question asked at least twice. A bundled Node.js script checks the question set and can sample model APIs for trend data when keys are present, kept apart from the manual results. Mentions, recommendations and citations are counted separately, errors are excluded rather than counted as misses, and a dated retest compares runs question by question.

Capabilities

  • Build a set of realistic test questions
  • Record how AI assistants mention or cite a brand
  • Sample model APIs for trend data
  • Diagnose why competitors are named instead
  • Retest and compare runs