If Tarela is the only live product, why does the site list three more?
The short answer is that applied intelligence, at the level we are building, is a portfolio problem, not a feature problem. A flagship implies one large product that everything else orbits. That is the right shape for some kinds of company. It is the wrong shape for what we do.
What each product line is for
- Tarela — customer-facing receptionist for WhatsApp and Instagram. Live.
- MemTask — persistent memory infrastructure for long-running autonomous tasks. In development.
- AgentPie — composable runtime for governed, multi-agent systems. In development.
- FashionTalk — conversational intelligence for fashion brands. In development.
The four share an architecture philosophy — clear product boundaries, dependable systems, practical user value — but they do not share a codebase. We are not building an "AI platform" that wraps a single big model. Each product earns its own place in a real workflow.
What we are optimising for
A small number of products that each earn their place by being the right shape for the workload. A salon owner needs a different shape than a research team running agents over months. Tarela is conversation-shaped. MemTask is memory-shaped. AgentPie is runtime-shaped. The workloads are different, so the products are different.
If you are evaluating NOVATOR.AI as a counterparty, the simplest entry point is Tarela. The other three will surface in stages through 2026 and 2027.