About
We build AI that takes its time.
Orchid is a research company making language models that read closely, reason carefully and are honest about what they don’t know.
Most AI is built to sound certain. We think the more useful thing is to be careful — to read everything, to show the work, and to say so when the evidence runs out.
We started Orchid after watching brilliant analysts spend their days double-checking confident machines. Our models are trained to slow down where it matters, cite what they rely on, and leave the final call with the people who know the work best.
Read everything
We don’t skim. Every page, footnote and appendix is part of the answer — or the reason there isn’t one yet.
Show the work
Citations, traces and confidence aren’t features we add later. They are how we decide a response is finished.
Leave room for doubt
A careful “I don’t know” beats a confident guess. We train for honesty first and fluency second.
2023
Founded in Lisbon by researchers who wanted slower, better answers.
52
People across 14 countries, most of them former researchers or editors.
$64M
Raised from investors who care about trust as much as speed.
1.2B
Tokens read every day by teams who need to get it right.
Who it’s for
Built
for
From contracts to clinical studies, Orchid is trusted wherever an answer has to hold up in the room.
Team
The people behind the model.
Researchers, engineers, editors and designers who would rather be right than first.
Elena Duarte
Co-founder & CEO
Lisbon
Kenji Watanabe
Co-founder & Chief Scientist
Tokyo
Mira Castellanos
Research Lead
Madrid
Theo Lindgren
Head of Product
Stockholm
Amaka Eze
Engineering Lead
Lagos
Jonah Mbeki
Developer Relations
Cape Town
Lucía Ferrer
Design Lead
Barcelona
Sam Okoro
Trust & Safety
London
Join a small team doing careful work.
We’re hiring across research, engineering and design — remote-first, with studios in Lisbon and London.