Why I built an AI assistant that doesn't know who you are

Dev.to / 4/3/2026

💬 OpinionDeveloper Stack & InfrastructureSignals & Early TrendsIdeas & Deep AnalysisTools & Practical Usage

Key Points

  • The author criticizes mainstream AI assistants for requesting personal identifiers, storing conversations, and tying usage data to user identities for model improvement.
  • They introduce NOLO, an AI assistant that assigns a random anonymous ID, avoids sign-up, and does not store conversations or keep logs on the server.
  • NOLO still provides access to powerful AI features such as real-time web search and image understanding, with privacy being the primary differentiator.
  • The implementation uses a simple stack: vanilla JavaScript frontend, Vercel serverless functions, Groq for faster inference, and Supabase solely for credit tracking (not conversation data).
  • The project is currently early-stage—one month in with no users yet—and plans to seek traction via Product Hunt, framing distribution as the main challenge.

Every AI assistant on the market asks for the same things: your email, your name, sometimes your phone number. Then they store your conversations, analyze your usage, and use your data to improve their models.

I get why they do it. Accounts mean retention metrics. Data means better models. It makes business sense.

But it also means that every question you ask an AI is being logged somewhere, tied to your identity, and kept indefinitely.

I wanted something different. So I built NOLO.

How it works

When you open NOLO, you get a random anonymous ID. No email. No signup. No personal information of any kind. Your conversations are not stored on our servers. There are no logs.

You still get access to powerful AI models, real-time web search, and image understanding. The privacy is the only thing that changes.

The technical side

The stack is simple: vanilla JavaScript on the frontend, Node.js serverless functions on Vercel, Groq for inference, and Supabase only for credit tracking (no conversation data, ever).

I chose Groq because the inference speed is genuinely faster than anything else I tested, which matters a lot in a real-time chat interface.

Where I am now

One month in. Zero users. Launching on Product Hunt this Tuesday.

The product works. Distribution is the hard part now.

If you care about privacy, I'd love for you to try it and tell me what you think.