Audit Engine
The audit engine queries 7 AI providers. Free audits use Gemini (20 prompts). Pro audits query all 7 engines with ~119 prompts including category and competitor queries.
- Gemini (Google) — via Gemini API
- ChatGPT (OpenAI) — via GPT-4o API
- Claude (Anthropic) — via Claude API
- Grok (xAI) — via Grok API
- Perplexity — via Perplexity API
- DeepSeek — via DeepSeek API
- Meta AI (Llama) — via NVIDIA NIM
All queries run at temperature 0 for deterministic, reproducible results. Each engine receives multiple prompt types: branded, discovery, comparative, .
Prompt Design
Branded prompts ask engines directly about the target brand. Discovery prompts ask for recommendations without mentioning the brand. Comparative prompts test head-to-head against competitors. The combination reveals:
- Whether engines know the brand exists
- Whether engines recommend the brand organically
- How the brand is described (sentiment, accuracy)
- Which competitors appear instead
- What sources engines cite
- How visibility varies across query types
Additional Analysis
Beyond AI engine queries, each audit includes parallel technical checks:
- SEO vs GEO comparison — Tavily web search as SEO proxy vs AI engine score
- AI crawler access — robots.txt analysis for 13 AI crawlers + llms.txt check
- Technical health — PageSpeed Insights (performance, SEO, accessibility)
- Authority sources — Wikipedia, Crunchbase, GitHub, LinkedIn presence
- Search demand — Google Autocomplete suggestions + People Also Ask
- Content AI-friendliness — On-page analysis for headings, FAQ, schema, entities (Pro)