Relari
Join Us

Robot dexterity, learned from human biomechanics.

Human dexterity is guided by muscles that continuously regulate contact. Relari brings that biomechanical signal to robotics so machines can learn more capable manipulation.

Backed by

  • Y Combinator
  • General Catalyst
  • Soma Capital
  • 468 Capital
  • Amino Capital
  • E14
  • Twenty Two Ventures
  • SBXi
  • Decacorn Capital
  • The Graduate Syndicate
  • Phoenix Investment Club
  • SID Venture Partners

EMG captures force signals hidden in videos.

A hand can look still while its muscles continuously regulate force, contact, and stability. Surface electromyography (EMG) captures the muscle activity behind those adjustments. Once decoded, this biomechanical structure can be transferred across robot embodiments, enabling richer, more human-like physical interaction.

Human demonstration
Robot execution

Measured Human Effort

EMG captures force timing, magnitude, and coordination.

Precise Robot Control

A force-aware control policy transfers that structure to the robot.

Non-invasive.
Easy to scale.
Rich in intent.

One lightweight armband adds interaction information to natural human demonstrations—without instrumented objects, gloves, or robot hardware.

01Grip force distribution
EMG patterns reveal how effort is distributed across fingers.
02Stiffness
EMG reveals muscle co-contraction as the hand braces or yields.
03Anticipation
EMG captures muscle activity before contact—and the force being prepared.
04Reflex
EMG captures rapid responses to slips and mistakes.

Force supervision, at scale

Force-labeled demonstrations are tied to a robot, operator, and lab. We study whether wearable EMG can move that supervision to people—across more tasks and environments.

A foundation model built from human dexterity.

Biomechanically rich demonstrations at scale enable us to train more capable, force-aware models for general-purpose robotic manipulation.

Our work sits at the intersection of robot learning, human biomechanics, and hardware. We’re looking for people with deep expertise, intellectual range, and a desire to explore problems that don’t fit neatly within one discipline.

Open positions