FAQ

What kinds of data does pdata handle?

Structured data and time series (Anu, Bhasa) and video (Chaya). Support for other data types is driven by pilot demand — talk to us if yours is different.

Do I ever need to decompress?

No. Querying, search, and forecasting all run against the compressed representation directly. There is no unpack step in the workflow.

What compression ratio should I expect?

Up to 1000x on real-world workloads, but it depends on your data. That's why pilots start with a sample of your data — we report the ratio and query behaviour you'd actually get, not a benchmark figure.

Where does it run?

At the edge, on-prem, or in the cloud. Compressing close to the source is a common deployment for video and sensor workloads, since it removes bandwidth and onboard storage as the limiting factor.

Is pdata in production anywhere?

Yes — live pilots run on UK electricity market data today, and first industrial pilots are underway in other verticals.

How do I try it?

Request a pilot. The process is described in Getting started.