Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.vangrid.io/llms.txt

Use this file to discover all available pages before exploring further.

Vangrid’s privacy model is architectural, not just a policy. Raw sensor data is processed directly on the edge node that captured it — your API calls never receive raw sensor streams and raw data never traverses the network. What reaches your application is derived, aggregated ground truth with a cryptographic provenance record.

How edge-computed privacy works

Every Vangrid edge node runs a local compute layer that:
  1. Captures multi-view observations from the physical environment
  2. Processes those observations locally into structured ground truth (geometry, classification, confidence)
  3. Signs the derived output with a node-specific cryptographic key
  4. Transmits only the signed, derived record — never the raw sensor stream
Raw inputs (images, LiDAR scans, RF signals) are discarded at the edge after processing. They are never stored centrally and never included in API responses.
This means you cannot retrieve raw sensor imagery or unprocessed data through the Vangrid API. You receive verified, structured spatial observations only.

What you receive vs. what stays at the edge

Data typeAvailable via APIStays at edge
Geometry (location, shape)Yes
Ground truth scoreYes
Observation classificationYes
Provenance hashYes
Raw sensor imagesNoYes
Unprocessed point cloudsNoYes
Node identity / locationAggregated onlyPrecise location

Privacy for your queries

Your spatial queries are also handled with privacy in mind:
  • Query parameters (AOI geometry) are not persisted beyond the request lifecycle
  • Results are scoped strictly to your API key’s authorized regions
  • No query history is shared across customers
If your use case requires data residency guarantees (e.g., data must not leave a specific country), contact hello@vangrid.io to discuss sovereign deployment options.