Open Source Robotics

A curated collection of high-quality open source robotics projects, tools, and software to propel the robotics community forward.

ORCA Hand

Open-source tendon-driven robotic hands with tactile sensing from ETH Zurich

License

  • orcahand_hardware (CAD, STL, 3MF files and its code): CC BY 4.0
  • orca_core and orcahand_description: MIT

Licensed by repository. orcahand_hardware (CAD, STL, 3MF files and its source code) is CC BY 4.0, copyright ORCA Dexterity, Inc.: free to share and adapt, even commercially, with credit, and no patent or trademark rights are granted. orca_core and orcahand_description are MIT. Checked 2 Nov 2026.

Overview

ORCA Hand is an open-source, tendon-driven dexterous robotic hand with 17 degrees of freedom, developed from ETH Zurich's Soft Robotics Lab. The current V2 release publishes print files, CAD and a motor connector board, with an assembly tutorial video and software on orca_core. The paper describing V1 reports a material cost below 2,000 CHF and a hold of up to 10.5 kg using all four fingers.

ORCA Hand is an open-source, 17-DOF tendon-driven dexterous hand. V2, the current release, publishes Bambu Lab print plates and STL and STEP files in orcahand_hardware, a motor connector board, and URDF and MJCF models for simulation; upstream publishes no written assembly manual, and instead offers a 92-minute tutorial video with an addendum. V1, described in the paper, used a different electronics and calibration setup, and the guide keeps it in a section of its own.

Why it matters

Commercial dexterous robot hands cost from around 15,000 CHF (Allegro Hand) to over 100,000 CHF (Shadow Hand) and are closed-source, creating a massive barrier for manipulation research. ORCA Hand democratizes dexterous manipulation by providing a reliable alternative, with a material cost below 2,000 CHF, and full open-source hardware designs, simulation environments, and research-grade tactile sensing. It enables labs worldwide to conduct sim-to-real transfer, imitation learning, and reinforcement learning research on dexterous hands without prohibitive hardware costs.

Problem it solves

Eliminates the cost and accessibility barrier for dexterous manipulation research. Before ORCA, researchers either spent around 15,000 CHF (Allegro Hand) to over 100,000 CHF (Shadow Hand) on commercial hands or built unreliable prototypes. ORCA provides a proven, affordable platform with simulation support for sim-to-real transfer.

Key takeaways

  • 17-DOF anthropomorphic hand with human-comparable grasp strength (up to 35N fingertip force)
  • Fully 3D-printable with under 8-hour assembly time and a material cost below 2,000 CHF
  • Complete digital twin ecosystem: URDF, MJCF, MuJoCo simulation, Isaac Lab support
  • Tested 10,000+ continuous cycles — designed for uninterrupted policy learning
  • Integrated tactile sensing with off-the-shelf barometric pressure sensors

Who it is for

  • Robotics researchers working on dexterous manipulation
  • Labs exploring sim-to-real transfer for hand manipulation policies
  • Students and educators in robotic grasping and manipulation courses
  • Hobbyists and makers interested in advanced robotic hands

Limitations

  • Tendon-driven design requires periodic tensioning and maintenance
  • Maximum 17-DOF — fewer than a human hand (27 DOF)
  • Upstream's V2 assembly is published as a video and an addendum, with no written manual.

This website requires JavaScript to function properly. Please enable JavaScript to explore our collection of open source robotics projects.