Street-level imagery
Privacy-blurred imagery from every drive, organised by map hex and refreshed as contributors pass through.
■ [ 03 · Map & Physical AI ]
Fresh ground truth for maps and machines.
Street-level imagery and map features from the network, refreshed in days, not years — for mapping, autonomy and physical AI.
[ What the network produces ]
Detected road features
Signs, lane markings, traffic lights and other features detected in the imagery and placed on the map.
The long tail of driving
Real-world driving events — the rare moments simulations can't produce — for training the models that drive.
[ Priced in Map Credits ]
A fixed dollar price, paid by burning HONEY.
Every Map Credit costs $0.0075. The Foundation sets how many credits each product needs, so prices stay stable while the HONEY required moves with the market.
How burn and mint works →| Product | Map Credits | USD |
|---|---|---|
| Imagery | ≈ 50 per km of road, per week | ≈ $0.375 / km / week |
| Map features | 1 per feature, per hex | $0.0075 |
Wholesale credit requirements as published in the HONEY docs, May 2025.
[ Why freshness matters ]
The road as it looks this week, not last year.
A
Autonomy & robotics
Models that act in the world need to know what changed — construction, new signs, new lanes.
B
Mapping & navigation
Maps stay accurate when they are refreshed by the people driving the roads every day.
C
Training data
Real event video captures what simulations can't — the rare, messy moments that teach driving models.