GEO Audit Tool - FastAPI Backend

A FastAPI backend for Generative Engine Optimization (GEO) auditing of educational course companies.

Features

• AI Answer Audit: Analyzes Perplexity AI answers for brand and competitor mentions
• Web Evidence Collection: Gathers data from web sources using Tavily API
• Fact-Check Pass: Verifies factual claims about the brand
• Reasoning Analysis: Uses CrewAI with Gemini to generate explanations and recommendations
• Aggregate Scoring: Computes brand mention rates, competitor share, hallucination rates

Setup

1. Install dependencies:

pip install -r requirements.txt

2. Set environment variables (optional, defaults provided):

export PERPLEXITY_API_KEY="your-key"
export TAVILY_API_KEY="your-key"
export GEMINI_API_KEY="your-key"

3. Run the server:

python main.py
# or
uvicorn main:app --reload

API Endpoints

GET /docs

Interactive API documentation (Swagger UI).

GET /health

Health check endpoint.

{
  "status": "ok"
}

POST /audit

Run a GEO audit for a brand.

Request:

{
  "brand_name": "Coursera",
  "brand_url": "https://www.coursera.org",
  "queries": [
    "best online courses for data science",
    "top educational platforms"
  ],
  "competitors": ["edX", "Udemy"]
}

Response: See the API specification in the code for the full response structure.

Architecture

• main.py: FastAPI application and endpoints
• services/perplexity_service.py: Perplexity API integration
• services/tavily_service.py: Tavily search and extract API integration
• services/crewai_service.py: CrewAI agent with Gemini for reasoning
• services/audit_processor.py: Main audit workflow orchestration