Your images are already free. Your index isn't.
Images leave your modalities pure: DICOM is a standard, and it works. What locks an imaging archive is the proprietary index built on top of it, and the closed orchestrator around it. Polnor rebuilds that layer as yours: a neutral, SQL-queryable index over standard tags, on your own bucket, with your AI reading and writing at home.
The format was never the problem.
Three things keep imaging data captive, and none of them is DICOM itself.
The proprietary index
PACS and VNA vendors index your studies with in-house tags and a closed database. Your images are standard; your ability to find and select them is not.
The closed orchestrator
Routing, prefetching, worklists: the logic that moves images lives inside the vendor's box. Every new use case becomes a vendor project.
Locked-in AI results
Vendor "AI partner" programs run models inside their walls, and keep the outputs there too. Your own predictions end up as screenshots, not data.
One standard gesture: add a destination.
Polnor is a first-class DICOM node on your own infrastructure. Your PACS admin adds it as a destination, the same gesture as any archive, and everything downstream builds itself.
Point your modalities or router at it
The Polnor receiver accepts C-STORE (DIMSE) and STOW-RS (DICOMweb), with calling-AE and IP allowlists. It runs as a container on your own compute, inside your network.
Images land pure, in your bucket
Every instance is written to your object storage exactly as sent, byte-for-byte Part-10. No transcoding, no proprietary wrapper, no copy on our side, ever.
The index builds itself
Standard tags (patient, study, series, modality, body part, station…) stream into an Iceberg table, imaging.dicom.instances. An existing store? A backfill scans it headers-only.
Your AI reads and writes at home
Select series in SQL, join with your cohorts and clinical data, train and serve models on your compute, and store predictions as SQL tables you own, next to the images they describe.
Not a PACS. Not a VNA. The neutral layer your AI was missing.
Your PACS remains the diagnostic viewer, and your archive remains the archive. Polnor is the layer that makes those images addressable by analytics and AI: an index you own, on data that never left you. Joinable with the rest of your warehouse, from the same catalog, under the same governance and access log.
An index you can query
« Every chest CT of my heart-failure cohort, 2020-2026 » is one SQL query over standard tags, not a vendor ticket.
Sovereign by construction
Pixels never transit through the control plane. Receiver, bucket, index, and models all live in your cloud, on an HDS-certified host.
Aligned with where France is going
Neutral archives and shared imaging access (DRIM-M) all point the same way: images and their index decoupled from any single vendor.
From a C-STORE push to a SQL row, in your cloud.
Book a 30-minute demo: we declare a receiver, push studies at it, and query the index — live, end to end, on infrastructure like yours.