We make machines sound like they mean it

Flabbergasted is a small team of engineers, linguists and sound people building voices that listen first and answer in the time it takes to breathe.

Work with us
Two colleagues in turtlenecks at a white desk with laptops, in front of a striped wall
Two people sitting on cream armchairs against a pink backdrop
A smiling woman working on a laptop while sitting on pink concrete steps
The team posing together on wooden chairs in a pink studio with bookshelves

2022

Founded in Stockholm

38

People across four cities

120

Languages and accents

4B

Characters voiced every month

Every voice product we tried sounded fine in a demo and fell apart in a real conversation. Pauses ran long, names came out wrong and the voice never changed its tone, whether it was booking a table or cancelling a flight.

So we built the engine we wanted: streaming first, tuned for the half second where a person decides if they are talking to something that listens. Everything else, from cloning to the Voice Lab, grew out of that one obsession.

Today teams use Flabbergasted to read audiobooks, answer support lines and give games characters that talk back. We still measure every release in milliseconds, and we still call the phone tree that started it all, just to check.

What we hold ourselves to

Three rules that settle most arguments in the office.

Consent first

No voice gets cloned without a recorded yes from the person it belongs to. Every clone carries a watermark.

Latency is a feature

If a reply takes longer than a breath, it feels like a machine. We ship nothing that makes us slower.

Ship the boring parts

Billing, logs, retries and docs get the same care as the voices. That is what makes it work in production.

The people behind the voices

Linguists, audio engineers and a few people who just really like talking.

Meet the team
Portrait of Marco Bianchi

Marco Bianchi

Head of Speech Research

Leads the team training our voice models, with a soft spot for prosody and a hard line on anything that sounds robotic.

Portrait of Maya Lombardi

Maya Lombardi

Voice Content Writer

Writes the scripts in our demos and docs, and teaches customers how to write for the ear instead of the eye.

Backed by people who bet on sound

$42M raised from funds and founders who have shipped audio, telephony and developer tools themselves.

Halden Ventures

Series A lead

Rocca Partners

Series A

Nordfjell Capital

Seed lead

Vesta Seed

Pre-seed

Arvo Fund

Seed

Lindqvist & Co

Angel