Pinned Tweet
We raised $7.5M to solve identity and fraud on the internet
Didit will be powering every application
We just raised $7.5M to build the infrastructure for identity and fraud.
The round is backed by @ycombinator, Orange Collective, Rebel Fund, Pioneer Fund, and other incredible VCs and angels.
AI is making fraud incredibly cheap to create: fake identities, fake businesses, fake transactions, all generated at internet scale. We believe every app will soon need trust infrastructure the same way apps today rely on payments or cloud hosting.
The future of the internet is verified. Didit (@didit) is building it.
How does a team of fewer than 20 serve 3,000+ customers?
Our CTO @_arosasg explains the setup behind Didit step by step: what we automated first, where the bottleneck moved each time, and the rules we follow.
Your team writes a rule, tests it on real traffic, and switches it on once they're confident. From then on, it reviews each transaction as it arrives and acts exactly as you told it to.
Live now in the Business Console, the API and Didit's MCP server.
didit.me/products/transactio…
We are next to the most ambitious building in the world. Started in 1882, it's still under construction, and the whole planet comes to see it. Gaudí used to say his client was in no hurry.
Ours are. In 2023 Didit was two brothers and two laptops. Today more than 3,000 companies run on it.
Last night it was our turn to have visitors. It poured, and 80 people from the startup scene still came to see us.
What a night, Barcelona. Didit was born in this city, and we like what it’s turning into.
Thank you all for coming. More on each talk next week.
"Everything humans have built runs on trust, but that belief is breaking"
That's how our founders @albertorosasg and @_arosasg begin the letter on why we build Didit.
didit.me/manifesto/
Proving who you are will be required for almost everything online. Whoever controls identity controls access to the internet.
Someone will build this layer of the internet, it matters who.
What we’re committed to:
→ Security: One system, constantly testing its own infrastructure. We ship defenses as threats appear and invite independent auditors to verify our work.
→ Accessibility: Free to start, priced for a two-person startup, and built to protect everyone from early-stage teams to global banks. Available in 220+ countries in any device, with a single prompt to integrate into your coding agent.
→ Privacy: Every check requests only what it truly needs.
Let's bring trust back.
Most companies keep how they ship to themselves. Our CTO posts the numbers.
This Thursday in our event in Barcelona, he'll show the rest: how we became an AI-native company, live.
Last month I shared that we merged nearly 2,300 PRs in 30 days at @didit.
Here is September.
1,997 pull requests merged.
89% of reviewed PRs were approved before merge.
Our AI review loop ran 1,965 review rounds across 836 PRs.
467 passed on the first round.
312 were sent back with changes, fixed, and approved on a later round.
"Should I raise in SF or in Europe?"
Bigger checks and higher valuations in the US. Capital efficiency and talent in Europe.
This Thursday two founders and one investor answer it from both sides of the table: @elisenda_bou (Cala), @borrellpepe (Masia) and @albertorosasg (Didit).
Barcelona, 6:30 pm
Save your spot: events.didit.me/e/ai-era
Didit retweeted
We just open-sourced Didit Events: an event app (think Luma) built around real, verified people.
You request to join, a selfie checks you're human, the host approves you, and at the door a camera recognises you (if you opted in) and checks you in. No tickets, no QR codes, and the host never sees your selfie.
Everything runs on Didit (@didit): liveness, optional ID checks, device + IP signals, duplicate faces, blocklists, investigations, and face search at the door.
I'm obsessed with identity-native apps. Identity is becoming programmable, and we'll open-source many more of these.
Repo: github.com/didit-protocol/di…
Didit retweeted
Things are going to get very strange, not only your grandma won’t be able to tell if it’s AI generated or not, but also you. We built @didit to solve this
Introducing Griffin, the first model to pass the video Turing test.
48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.
It’s the first Human Interaction Model (HIM).
Readers added context they thought people might want to know
The 48% figure and "video Turing test" claim are from Tavus's own study of 54 one-minute calls, not independently verified or using a standard protocol. Griffin-Lite leads NVIDIA's VideoFDB benchmark on their public leaderboard.
cellcog.ai/blog/tavus-gri…
research.nvidia.com/labs/amri/proj…
tech-ish.com/2026/10/02/tav…
Announcing Didit's first event in Barcelona.
Join us for the opening of our new office, alongside the people building the city’s AI startup scene.
We’ll talk about AI-native companies and fundraising in SF vs EU, then celebrate with a proper housewarming party.
October 8th · 6:30 PM
Barcelona
Building, investing, or curious about AI? Come by
Introducing Didit KYB, know-your-business for companies and the humans behind them.
A company is never just a company. It is a registry file, a list of owners, and a set of real people who each have to be checked.
One session does all three and files the result on the business record.
What has to change.
A document image no longer proves who is behind a transaction.
Checks have to verify a real person is on a real camera, right now, and hold up against replay and injection attacks.
And the industry has to stop treating permanent retention of identity data as normal.
Collect less. Retain less. Verify better.
