About Thyra AI

AI assistants that people and companies can actually trust.

Thyra AI Corporation builds AI assistants that do real work (email, calendars, follow-ups) while keeping a person in charge of anything important. The company started as a Google Gemini 3 Hackathon project that grew into a real product. The founding team combines product and operations experience with AI research, all focused on one thing: help you can rely on, without giving up control.

Timeline

Our story

Corporate history from research through formation of Thyra AI Corporation.

1. Before February 2026

   Research foundation

   José completed his master's thesis, Optimizing Small Language Models for Resource-Constrained Environments, an adaptive query-routing framework for efficient on-device and low-resource deployments.
2. Early 2026

   The team comes together

   Farley had recently left a startup and was exploring what was next. He invited José to team up for the Google Gemini 3 Hackathon, combining product execution with José's research background.
3. February 4, 2026

   Joined the Google hackathon

   We entered the Gemini 3 Hackathon and started building in earnest.
4. February 10, 2026

   Submitted to the Gemini Hackathon

   We

   submitted

   our hackathon project: a working end-to-end prototype on the Gemini stack for what would become Thyra.
5. February 12, 2026

   From hackathon to product

   We decided the prototype was worth turning into a real company and product, not a one-week demo.
6. March 11, 2026

   Thyra AI incorporated

   Thyra AI Corporation was formed in Delaware, United States.
7. April 4, 2026

   DigitalOcean startup program

   Thyra joined DigitalOcean Hatch, with sponsorship through the startup program for cloud infrastructure and founder support.

Our mission

AI help you can supervise, for one person or a whole company.

AI is getting very good at doing work. What's missing is the part that makes it safe to hand work over: knowing what your assistant is allowed to do, being asked before anything important happens, and being able to see exactly what it did. That's what Thyra builds.

Why now

AI assistants are starting to do real work everywhere. What holds most people and companies back isn't how smart the AI is. It's trust. Without clear limits and a record of what happened, handing work to an AI feels risky, and cleaning up after it gets expensive.

Our focus

We don't try to do everything. Thyra focuses on three things: each assistant has clear permissions for the apps it can touch, you decide how it behaves beyond just "prompting" it, and a person is asked whenever a decision really matters, like sending money, contacting a customer, or anything with legal weight. It fits how families and companies already make decisions.

What we're building

A platform where anyone (an individual or an organization) can set up assistants the same reliable way every time: control over which tools and data each one can reach, safety rules that don't depend on wording a prompt just right, and approvals that go to the right person. More help, without more risk.

Where this is going

More and more everyday work will be done by AI assistants. The ones people keep will be the ones they can check, correct, and trust. Thyra is built for that future.