Track how ChatGPT, Gemini, Perplexity and Claude talk about your brand.

Overview DolomiteAI is a live SaaS platform (dolomiteai.com) that monitors how major AI assistants represent a brand in their generated answers. As more people ask AI tools for recommendations instead of searching Google, brands need to know whether they show up, and how. DolomiteAI measures that and turns it into clear next steps.

Multi-model monitoring. Brand visibility is tracked across ChatGPT, Gemini, Perplexity and Claude.

Automated monitoring engine. Each brand has its own cron schedule, run through Supabase pg_cron. I debugged and stabilized the engine, resolving four interconnected issues across the monitoring hooks and the daily cron job, and moved the schedule from daily to hourly.

Visibility scoring. I aligned the measurement methodology to binary scoring and wrote a client-facing document explaining the change.

Recommendation engine. A scoring-based "Visibility Opportunities" system that turns monitoring results into prioritized actions. It uses a three-part architecture: a template registry, a resolver and a UI component, so new recommendation types are easy to add.

Multitenant architecture. Brands and accounts are separated cleanly on a shared Next.js and Supabase foundation.

Staging environment. A separate staging site (staging.dolomiteai.com) on Vercel with a stable alias, so changes can be tested before reaching production.

Tech stack Next.js, React, TypeScript, Supabase, PostgreSQL, pg_cron, Vercel, and the Claude, Gemini and other AI model APIs.

Why it matters for your project This project shows the kind of work I do for clients: AI API integrations, scheduled background jobs, multitenant data design, and debugging a live product until it's reliable.