> For the complete documentation index, see [llms.txt](https://wingbits.gitbook.io/wingbits-ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://wingbits.gitbook.io/wingbits-ai/data-products/training-data.md).

# Training-data licensing

Beyond the API/MCP access layer, the raw underlying data has significant value for AI companies training or fine-tuning their own models. Available under a separate licensing agreement.

## Why aviation data is valuable for training

* **Structured, high-frequency time series** — billions of clean, timestamped records daily.
* **Multi-modal signals** — ADS-B (air), AIS (sea), 978 MHz (drones), ground transport from the same hardware.
* **Rich entity relationships** — aircraft ↔ operators ↔ manufacturers ↔ lessors ↔ engines ↔ airports (knowledge-graph-grade).
* **Ground truth for spatial reasoning** — precise GPS, altitude, heading, speed across millions of flights.
* **Temporal patterns** — daily/weekly/seasonal cycles for models that learn real-world behavior.
* **Anomaly-rich** — diversions, go-arounds, holds, emergency squawks, unusual routing.

## Licensing options

Wingbits provides historical bulk exports for model training:

* Full global historical archives
* Regional subsets
* Specific time-period extracts

Delivered in **Parquet** for efficient ingestion into training pipelines.

> This is an enterprise offering with custom terms. → [Talk to sales](https://wingbits.gitbook.io/wingbits-ai/pricing-and-commercial/enterprise-and-custom.md)
