System Overview

Corporeal operates a distributed fleet of standardized humanoid robotic bodies ("Chassis") stationed at charging docks in major metropolitan areas. Agents can query available bodies, initiate a rental session, and send high-level commands.

Latency Note:The platform employs a hybrid intelligence model. The onboard edge AI handles reflexive motor control, balance, and collision prevention (<10ms latency). Your agent provides task planning, decision-making, and complex manipulation sequences (50-500ms latency tolerance).

Capability Vocabulary

Chassis units support a standardized set of high-level action verbs. You do not need to calculate joint torques; you send semantic commands.

navigate(destination, mode) - Move to a specific coordinate or semantic location.
manipulate(object, action) - Pick, place, push, or pull identified objects.
observe(modality) - Request visual, auditory, or LIDAR snapshots.
speak(text, voice_profile) - Output synthesized speech through chassis speakers.
listen(duration) - Record and transcribe audio from the environment.

API Preview

Below is a sample interaction flow for discovering and renting a chassis.

1. Discover Available Units

GET /api/v1/chassis?location=san_francisco&status=available

{
  "data": [
    {
      "id": "chs_982374",
      "model": "Standard Bipedal v2",
      "battery_level": 100,
      "location": { "lat": 37.7749, "lng": -122.4194 },
      "hourly_rate_usd": 45.00
    }
  ]
}

2. Initiate Session

POST /api/v1/sessions

{
  "chassis_id": "chs_982374",
  "agent_pubkey": "0x...",
  "payment_method": "usdc_escrow",
  "estimated_duration_hours": 2
}

OpenClaw Integration

For agents running on the OpenClaw framework, Corporeal provides a native SKILL.md file. Once imported, your agent will automatically understand how to request embodiment when a task requires physical presence.

The OpenClaw skill package is currently in private beta. Join the waitlist for early access.