Mimic Robotics

Frontier AI for
robot dexterity

We build video-action models for general purpose robotics. Our robots intuitively learn new skills to autonomously scale tedious operations from manufacturing to logistics.
contactabout

Automating unstructured industrial work

Our end-to-end learning approach for dexterous manipulation solves arbitrary physical work with minimal fine-tuning to automate unstructured tasks with changing positions, objects, soft bodies, and cables, previously impossible for conventional automation.

01

Assembly

Complex manual assembly steps and kitting processes in manufacturing.

02

Packaging & Sorting

In- and out-of-the-box handling processes for many irregular objects.

03

Complex Manual Tasks

Dexterous manipulation tasks in unstructured environments.

Video-Action Models to scale robot learning beyond VLAs

We are pioneering video and world models to leverage large-scale native video pre-training with rich physics and dynamic representations for the generation of robot actions.
technology

Recent highlights

6/1/26Blog

Video-Action Models: Are video model backbones the future of VLAs?

This blog post is about mimic-video, our latest mimic release in which we instantiate a new class of Video-Action Models (VAM), grounding robotic policies in pretrained video models. We argue that video model backbones can be a much more natural choice for robotics foundation model pre-training compared to VLM backbones.

Video-Action Models: Are video model backbones the future of VLAs?
17/12/25arXiv

mimic-video: Video-Action Models

Generalizable robot control beyond VLAs with pretrained Internet-scale video model backbones.

mimic-video: Video-Action Models
17/7/25CoRL

Diffusion for Cross-Embodiment Manipulation

A novel diffusion-based framework to align action spaces across human hands, mimic’s humanoid robotic hands, and other grippers. By training a unified latent action space via contrastive learning, LAD enables a single policy to control diverse robot embodiments, delivering up to a ~25 % lift in manipulation success and unlocking scalable skill transfer across robotic platforms.

Diffusion for Cross-Embodiment Manipulation
1/7/25Blog

Form Follows Data

One of the first questions we are always asked at mimic is: "Why the human hand form factor?" It's a fair question. Indeed, it is not obvious why we would want to focus on this specific morphology, when the design space of grippers allows for so many alternatives, from simpler two finger grippers, to hands more capable than the human, for example with 6 fingers. Here I want to lay out the design philosophy guiding our quest towards solving general-purpose robotic dexterity and what I think the future of the field will look like. As the title already suggests, the reason isn't purely functional: it’s a choice driven by data availability.

Form Follows Data
13/6/25CoRL

mimic-one: a Scalable Model Recipe

A next-generation diffusion-based control architecture built around a newly designed 16‑DoF tendon‑driven humanoid hand. Using glove and VR teleoperation data, MIMIC‑ONE achieves up to 93.3 % task success (even on unseen tasks) by learning smooth, self-correcting motion policies from sensory inputs alone.

mimic-one: a Scalable Model Recipe
8/4/25arXiv

MAPLE: Priors Learned From Videos

In collaboration with ETH Zürich and Microsoft Research, MAPLE introduces a new method to train real-world dexterous robotic hands by learning from large-scale video of human hand interactions. It shows how manipulation priors extracted from egocentric footage enable robots to grasp, manipulate, and generalize across varied tasks.

MAPLE: Priors Learned From Videos
24/10/24Blog

Robotics in the era of the Scaling Hypothesis

This piece is my attempt to collect my thoughts on the current landscape in AI and Robotics, how this informs our approach at mimic in our quest towards solving general-purpose robotic dexterity, and what I think the future of the field will look like.

Robotics in the era of the Scaling Hypothesis
Read More

We empower people to focus on what truly matters

Solving the hardest tech challenges takes a team that’s obsessed with making it real. At mimic, we brought together top minds in AI, robotics, and engineering to build the leading team for frontier physical AI in Europe.
about

Always the right robot for your application

Our robot stations are available with one or two arms on a flexible stationary table or autonomous mobile platform. This platform agnostic approach makes sure we always have the robot best tailored to your specific use-case.
contact

Unlock automation across your operations

At mimic, you are at the centre of all we build. We are committed to a fast, transparent process to turn your operational challenges into deployed, fully autonomous solutions.
contact

Intro call

In a non-binding 20 min call, we explore your current workflow challenges and identify where our automation solutions can drive the most value.

Use-case analysis

Our automation experts visit your facility to understand your unique needs, environments, and processes. We craft a custom proposal aligned with your operational priorities, timeline, and budget.

Deploy & scale

We install and integrate a mimic robot station, tailored and fine-tuned for your application, directly into your workflow. Rapidly adapt your robot to new use cases and scale effortlessly.