What We're Building — Ricordo | A Personal Travel Agent That Learns Your Taste
What we're building

The best places have already been discovered.
They just haven't been revealed to you.

Where to eat, what to do, where to go — we make these decisions constantly, and they're some of the biggest discretionary money we spend. Yet finding the best answer still means digging through hundreds of sites.

277 web pages viewed per trip
5+ hrs planning a single week away
more research time vs. 2016

Source: Travel Market Report; Expedia path-to-purchase research

The opportunity

Travel tech spent 25 years booking flights and hotels. The real money is what we spend every day.

The industry built itself around the once-a-year trip — the flight, the hotel, the rental car. But the spending that actually shapes our lives is weekly: the dinners, the outings, the neighbourhood finds. That spending has never had software on its side.

And the data underneath it is unusual. Place knowledge is the web's densest public data — and its most perishable. Openings, closings, seasons, chef changes, the room that's lost its magic. This is exactly where persistent AI agents will outperform today's search box: not answering once, but watching continuously.

The hard part

Getting this right takes deep knowledge of four things

1

You

Your taste, your history, and the people you're with.

2

Your contexts

Where you go, what you do. Date night, the Lisbon trip, the cottage hunt, the best egg sandwich in the city.

3

The places within them

Restaurants, stores, streets, shows, stays.

4

A changing world

Openings, closings, seasons, events. The map never sits still.

That's Ricordo

Your contexts, watched.
Your places, revealed.
And why.

Your memory

A personal knowledge graph: tastes, companions, contexts. It compounds with every interaction.

Your agents

AI agents continuously discovering and evaluating for you — surfacing the right option at the right moment.

On your side

No bookings. No commissions. No paid placement. You pay a subscription, and it answers honestly.

Where we start

First: the urban outing-goer

Ricordo begins with the person deciding where to spend Friday night — a weekly use case, not a once-a-year trip. The same memory and the same agents then travel with you, so the guide that knows your city already knows you when you land somewhere new.

Built, not imagined

A working system, in daily use

Ricordo runs today as a working multi-agent system with persistent memory — rebuilt five times, each rebuild killing a wrong assumption and uncovering a higher-impact use. It's in daily use, at home and away: better nights, experiences that would have been missed, fewer disappointments. And a growing waitlist of people who want the same.