What is pdata?
pdata is Precica's compression engine. It compresses data by orders of magnitude — up to 1000x on real-world workloads — and keeps it directly queryable: you search, recall, and forecast against the compressed representation itself, with no decompression step.
That single property removes the usual tradeoff between storing data and using it. Instead of downsampling history, pushing it to cold storage, or deleting it, you keep everything — hot.
The model family
pdata isn't one model, it's a family. Each model handles a different kind of data:
| Model | Data | Capabilities |
|---|---|---|
| Anu | Structured data | Compression |
| Bhasa | Structured data | Compression + forecasting |
| Chaya | Video | Compression + search |
Where it runs today
- Energy — live pilots forecasting UK electricity markets from compressed meter and market history.
- Physical AI — compressed, searchable memory for robots: recall where an object was last seen instead of re-scanning the room.
- Security — years of surveillance footage retained with every frame still searchable.
Explore the docs
-
How pilots work
pdata is deployed through guided pilots today. What to expect, step by step.
-
The model family
Anu, Bhasa, and Chaya — what each model does and when to use it.
-
Guides
Querying, forecasting, video search, and edge deployment — being written alongside the pilot programme.
-
FAQ
Data types, compression ratios, deployment options, and how to try it.
Questions the docs don't answer yet? Contact us.