mise

The layer every restaurant will run on. When your customers' AIs start booking tables and ordering dinner, mise is what they speak to.

Table mid-service
Client
Tiger Forest
Timeframe
2026

Nothing talks to anything

A venue runs on systems that don't talk to each other: bookings in one, the POS in another, stock and rota somewhere else. I learned that the hard way at QikServe, building more than 130 integrations. It was a problem when the customer was a person standing at a kiosk. It becomes urgent when the customer is an AI.

mise is an AI-native rail that connects those fragmented systems and opens them, safely, to every agent: your customers' AIs, staff copilots, operations.

Everything in its place

The name is from mise en place: everything in its place before service. That is the job mise does for a venue's data.

Where it is

There is a proof of concept that pulls bookings, POS, stock and rota into one model an agent can read. Alongside it, conversations with operators and with possible cofounders.

Free to start

The model follows Vercel: free to start, then pay by volume of use. It is chosen for growth and for disruption. I want to get it into the hands of developers.

The hardest problem is adoption: getting operators and developers to build on it before agents are mainstream.