Full Spatial Intelligence just an API call away.
Vision Is
All You Need.
Ditch expensive LiDAR and complex sensor fusion. Give your robot full spatial awareness using only a standard video stream.
Built for Builders
Initialize the client, pipe in your camera stream via WebRTC, and receive navigation vectors in real time.
Read the Docs →import asyncio
import odyseus as od
from aiortc import RTCPeerConnection
from aiortc.contrib.media import MediaPlayer
import robot_driver
async def main():
client = od.Odyseus(api_key="sk_your_api_key_here")
robot = robot_driver.Chassis()
player = MediaPlayer("/dev/video0", format="v4l2")
pc = RTCPeerConnection()
camera_track = od.webrtc.LatestFrameTrack(player.video)
pc.addTrack(camera_track)
await client.connect_webrtc(pc)
while True:
frame = await camera_track.recv()
result = await client.infer(frame.to_image())
robot.move(result["command"])
asyncio.run(main())Hardware Agnostic Cloud Platform
BUILT FOR REAL ROBOTS.
Live video, depth estimation, memory, movement, and cloud deployment in one robot stack.
| Capability | Odyseus | Vision API | Generic World Model |
|---|---|---|---|
| Live autonomy loop | ✓ Streams camera frames, runs inference, and returns robot actions in one loop. |
✕ | sim first |
| Monocular depth estimation | ✓ Reasons about distance without extra depth sensors. |
✕ | different job |
| Spatial memory | ✓ | ✕ | bounded |
| Robot movement commands | ✓ | ✕ | custom |
| No LiDAR / digital twin required | ✓ | image only | ✕ |
| SDK + cloud deployment | ✓ | api only | infra heavy |
Lower Hardware BOM
Remove the sensor burden. Odyseus enables autonomy on low-cost hardware via standard visual sensors.
Instant Autonomy
Abstract the physics. Go from unboxing to autonomous navigation in under 60 minutes.
Any Dev Team
Scale without robotics PhDs. Odyseus democratizes robotic autonomy.
One Model. Any Platform.
