About Simular

Building machine autonomy to unleash human potential.

If you’re reading this, you’re probably like us – a knowledge worker who spends well over 1,500 hours a year just clicking, typing and moving on the computer.

We're a research-driven company chasing one goal: full machine autonomy. Every product we ship moves us closer to a world where AI takes over digital labor, freeing people to develop and flourish.

Our journey

From a live demo to human-level agents.

Some of our milestones since our founding in August 2023.

Scroll to explore

2023
August 2023 · Company
Simular is founded
We set out to build accessible general intelligence and free people from digital labor.
December 2023 · Live demo
Our agent schedules a meeting live on stage
In front of a packed room in San Francisco, Simular's agent called a business and booked a meeting — end to end, on its own.
2024
2024 · Research
Earliest to work on OSWorld and ship computer-use agents
Agent S, our open framework for using a computer like a human, becomes one of the first computer-use agents anyone can run.
2025
April 2025 · Research
Best Paper at the ICLR Agentic AI for Science workshop
Agent S2 is recognized at ICLR 2025 for advancing autonomous agents in scientific work.
December 2025 · Research
Agent S becomes the first to reach human-level performance on OSWorld
72.6% on OSWorld — the first computer-use agent to beat the human baseline.
Today

2023 — todayCompany milestones

Our approach

Build autonomous systems in real life and bring more freedom to people

Real life

Autonomy needs a closed feedback loop. Continual learning in real life is where actions produce consequences that feed back in.

Full-stack

Autonomy means optimally free. We build not only the user-facing agent interface, but also the virtual machine that the agent runs on to free humans from devices.

Neurosymbolic

Autonomy needs accountability. Our neurosymbolic framework combines the exploratory power of LLMs with the deterministic reliability of code, balancing capability and reliability.

Equipe

Somos um pequeno grupo de cientistas e engenheiros da DeepMind, do Google e dos principais laboratórios de IA. Trabalhamos em sistemas de aprendizado por reforço de ponta, design de compiladores e robótica do mundo real — e agora estamos aplicando essa experiência para criar agentes que realmente funcionem.

Desenvolvido pela equipe de especialistas em IA da