Embodied AI
A continuous feed of verified spatial ground truth for robots and other physical agents.
The API is in early access and is not publicly available yet. Endpoint URLs, request examples and sample responses on this page are illustrative. Contact hello@utopiadata.net to request access.
Safe and effective operation for any embodied AI system (a humanoid robot, a drone swarm, a piece of autonomous industrial equipment) depends on holding a persistent, always current model of the physical world. That spatial cortex is what Utopia provides: a real time stream of verified ground truth observations that your system ingests to keep its world model in sync with reality.
What embodied AI needs from spatial data #
No world model outperforms its inputs. Utopia targets the three requirements that matter most: continuous updates, because the physical world never stops changing and your world model cannot fall behind it; verified observations, because a provenance_hash on every data point lets your system trust what it ingests; and high spatial density, because tactical urban coverage yields the node overlap needed to corroborate observations and drive out false positives.
Streaming spatial updates into your world model #
Subscribing to a geographic zone through the Utopia Streaming API gets you push updates whenever ground truth changes.
Define your zone
Describe the geographic polygon your system works inside. A tight area keeps both update volume and latency down.
Open the stream
Authenticate to the streaming endpoint with your API key. Events arrive as newline delimited JSON.
Update your world model
Parse the incoming events into your spatial representation, weighting each observation by its ground_truth_score.
Verify before acting
Where a decision is safety critical, check the provenance_hash first, then act on the observation only after it verifies.
import requests
import json
API_KEY = "your-api-key"
BASE_URL = "https://api.utopiadata.net/v1"
aoi = {
"type": "Polygon",
"coordinates": [[
[-122.4194, 37.7749],
[-122.4094, 37.7749],
[-122.4094, 37.7849],
[-122.4194, 37.7849],
[-122.4194, 37.7749]
]]
}
with requests.post(
f"{BASE_URL}/spatial/stream",
headers={"Authorization": f"Bearer {API_KEY}"},
json={"aoi": aoi, "min_ground_truth_score": 0.80},
stream=True
) as response:
for line in response.iter_lines():
if line:
event = json.loads(line)
update_world_model(event)Key data fields for embodied AI #
| Field | Type | Description |
|---|---|---|
| geometry | GeoJSON | Where the observed feature is and what shape it has |
| ground_truth_score | float, 0 to 1 | Confidence derived from multi node corroboration |
| timestamp | ISO 8601 | The moment the observation was captured |
| provenance_hash | string | Cryptographic evidence of where the data originated |
| node_count | integer | How many edge nodes contributed |
| observation_type | string | The observation's category, such as obstacle, occupancy or change |
A ground_truth_score floor of 0.75 suits real time navigation. Raise it to 0.90 or above for safety critical actions like grasping or human interaction.
Example streaming event #
{
"event_id": "evt_8f3a2c1d",
"timestamp": "2026-05-22T10:14:33.412Z",
"geometry": {
"type": "Point",
"coordinates": [-122.4150, 37.7800]
},
"observation_type": "obstacle",
"ground_truth_score": 0.94,
"node_count": 12,
"provenance_hash": "sha256:a3f8e2c1b4d7f9e0a2b5c8d1e4f7a0b3c6d9e2f5"
}Running embodied AI in production? Contact hello@utopiadata.net to talk through high frequency streaming rates and dedicated node allocation.