A twin that attends the meeting with your judgment
Notetakers capture what was said. We are building something different: a verified twin of you that joins the meeting, asks the questions you would ask, evaluates the conversation the way you would, and leaves a memory you can interrogate afterwards. “Would I have taken this deal? What did I miss?”
It builds on what already runs today: fidelity-scored twins of named, consenting experts, holding live conversations billed per minute.
Early access: become a design partner
First cohort: investors running founder meetings. Experts and operators welcome too.
Consent-first, by design
An AI that participates in meetings carries real obligations: recording consent, transparency, data protection. We are building for them rather than around them: the twin announces itself before it speaks, participants consent before it takes part, its identity is verified so nobody wonders who is behind the face, and every evaluation is decision-support for a human, never an automated decision about a person. That is the same consent-native approach the live twin platform already ships.
Common questions
- Is the meeting twin available today?
- No. It is in design, and we are selecting early design partners now. Registering interest is how you become one.
- How is it different from a notetaker?
- A notetaker answers "what was said". A meeting twin is being built to answer "what would I have concluded". It applies a specific, consenting person’s judgment and leaves a queryable memory of the meeting, not just a transcript.
- Is that even legal in the EU?
- That question is exactly why it is consent-first by design: the twin announces itself, participants consent before it takes part, identity is verified, and its evaluations are decision-support for a human, never an automated decision about a person. We are building with counsel review before any pilot.
- Who is it for first?
- Investor diligence: partners who take many founder meetings and want their own questions asked and their own frameworks applied consistently. Other verticals follow.
The judgment engine already answers questions in public. Try the live flagship twin: