Virtual · Physical · Teleoperated · Autonomous

Embodied Agency

Open Vessel explores the forms through which intelligent systems perceive, inhabit, and act within environments: virtual avatars, teleoperated machines, robotic bodies, simulated agents, and hybrid systems that distribute control across human, AI, and physical interfaces.

The central question is practical: how does intelligence become situated enough to act? Open Vessel treats embodiment as a continuum rather than a silo, connecting software intelligence, avatar presence, robotics hardware, motion capture, human input systems, simulation, and environmental feedback.

Open Vessel hero image: two masked humanoid avatars and a quadruped robotic form inside a geometric field

Project Entry Points

These three doors form the current public foundation. Nexus already runs as a public demo. Avatar Workbench remains staged for the next build. Embodiment Resources provides the expanding third-party engineering atlas.

Active Public Demo

Nexus Interface

A modular AI interface for experimenting with model providers, local SMOL variants, visible system prompts, voice output, and embodied mask presence. Nexus represents the current software intelligence layer of the Open Vessel stack.

Launch Nexus Demo

Active Public Demo

Avatar Workbench

A modular web based avatar facial mocap interface for developing microexpression resolution capable facial mocap layer for the Open Vessel stack.

Launch Avatar Workbench

Resource Atlas

Embodiment Resources

An expanding engineering atlas for robotics platforms, locomotion systems, actuators, sensors, controllers, simulation tools, datasets, open-source projects, manufacturers, and research references.

Browse Resources

Research Areas

Open Vessel organizes embodiment by function rather than by industry, allowing virtual bodies, physical robots, simulated agents, and human interface systems to share a common map.

Artificial IntelligenceModel providers, local agents, prompts, memory, tools, and decision systems.Active

Virtual AvatarsHumanoid and nonhuman digital bodies for presence, navigation, and interface testing.Demonstrator

Robotics HardwareHumanoids, quadrupeds, manipulators, drones, marine systems, and mobile platforms.Resource Atlas

Human InterfacesControllers, voice, eye tracking, EEG, EMG, exoskeletons, haptics, and accessibility devices.Prototype

Motion CaptureBody tracking, pose estimation, locomotion capture, animation transfer, and training data.Prototype

Facial CaptureExpression tracking, blendshapes, phonemes, mask animation, and embodied speech.Prototype

SimulationTraining environments, digital twins, synthetic data, reinforcement learning, and safety testing.Research

TelepresenceRemote embodiment, shared control, sensor feedback, camera rigs, and situated operation.Prototype

LocomotionWalking, rolling, crawling, swimming, flying, balancing, recovery, and terrain adaptation.Resource Atlas

ManipulationHands, grippers, arms, dexterity, grasping, tool use, and object interaction.Resource Atlas

Embodied AIAgents whose intelligence couples to sensors, actuators, memory, motion, and feedback.Research

Multi-Agent SystemsCollaborative bodies, swarms, shared worlds, field agents, and distributed action.Prototype

Open Vessel does not catalog robots alone. It studies the complete ecosystem required to instantiate intelligence into an embodied system.

Whether intelligence operates through a biological organism, a virtual avatar, a teleoperated robot, an autonomous machine, or a simulated agent, each form answers a related engineering question: how can perception, cognition, control, and feedback become coupled strongly enough to produce situated action?

For that reason, Open Vessel treats embodiment as a continuum. A face-capture mask, a VR avatar, a quadruped robot, a drone, a prosthetic hand, and a simulated humanoid controller all belong on the same map when viewed through the lens of agency entering an environment.

The current public page establishes the silo. The linked tools and resource atlas provide the first working surface. Future modules will expand this into a fuller Avatar Workbench, motion-capture pipeline, human-interface atlas, and embodied-AI laboratory.