I built a SaaS to measure AI visibility (GEO) — here’s how it works

Dev.to / 3/28/2026

💬 OpinionDeveloper Stack & InfrastructureSignals & Early TrendsTools & Practical Usage

Key Points

  • The article introduces GeoTracker, a SaaS web app designed to measure how AI systems “perceive” a company in AI-generated answers.
  • GeoTracker works by generating realistic user-style prompts, querying multiple AI models, and extracting brand mentions, competitor presence, and cited sources from the responses.
  • Its core features include multi-model analysis, manual and batch prompt testing, source extraction, a GEO recommendations engine, and PDF exports for client reporting.
  • The platform is built with React/Vite (frontend), FastAPI (backend), PostgreSQL (database), and Docker/Nginx on VPS, using secure cookie-based JWT authentication.
  • The author frames the product as addressing the shift from traditional SEO to “Generative Engine Optimization (GEO),” and outlines next steps like automated monitoring, scoring/trends, and a multi-client SaaS version.

AI is quickly becoming a new search layer.

People don’t just Google anymore — they ask ChatGPT, Claude, Gemini…

But here’s the problem:
Companies have no idea if they are visible in these answers.

So I built a tool to solve that.

Introducing GeoTracker

GeoTracker is a web app that analyzes how an AI perceives your company.

It answers questions like:

  • Is my company mentioned in AI responses?
  • Which competitors are shown instead?
  • What sources are used by the AI?
  • How can I improve my visibility?

How it works

The idea is simple:

  1. Generate real-world prompts (like users would ask)
  2. Send them to multiple AI models
  3. Analyze the responses
  4. Extract:
    • brand mentions
    • competitors
    • cited sources
  5. Generate actionable recommendations

Core features

  • Multi-model AI analysis (OpenAI, etc.)
  • Manual prompt testing
  • Batch testing (run multiple prompts at once)
  • Brand & competitor detection
  • Source extraction
  • GEO recommendations engine
  • PDF export for client reports

Tech stack

  • Frontend: React + Vite
  • Backend: FastAPI
  • Database: PostgreSQL
  • Infra: Docker + Nginx (VPS)
  • Auth: Secure cookie (JWT)

Why this matters

We are entering a new era of search.

SEO is no longer enough.

We now need to optimize for AI-generated answers:
GEO (Generative Engine Optimization)

If your company is not mentioned in AI responses, you are invisible.

Example use case

Prompt:

“Best suppliers of slip rings in aerospace”

Result:

  • Your company might not appear
  • Competitors dominate
  • Sources come from specific domains

That’s actionable data.

What’s next

I’m currently working on:

  • automated monitoring
  • scoring & trends
  • multi-client SaaS version
  • better recommendations engine

Final thoughts

This project started as an experiment, but it quickly turned into something bigger.

AI visibility is going to be a major challenge for companies in the next years.

And we’re just at the beginning.

If you’re working on SEO, AI, or similar tools, I’d love to connect

広告