The UPC's New Case Management System: What Changed for Data Access

On 8 July 2025, the Unified Patent Court switched on the first phase of a rebuilt UPC CMS — its Case Management System — developed in cooperation with the European Patent Office. For most users the headline is a smoother filing experience: software-based logins instead of physical smartcards, document uploads at any time, and a single case file in place of parallel case numbers. For anyone who works with UPC data the headline is different, and quieter: the new system modernises how cases go in, but changes almost nothing about what comes out. Decisions are still published as anonymised PDFs, pleadings still require a reasoned request, and there is still no public analytics layer, no full-text search of decisions, no bulk export, and no open data API. The registry tells you a decision exists. It still won't tell you who won — and that gap is precisely the one our database is built to fill.

What the new UPC CMS actually changed

The old Case Management System had a rough reputation. Live since the Court opened in June 2023, it drew sustained complaints from users and judges through its first operational year — a workflow-bound design that forced parallel case numbers and leaned on smartcards and USB dongles for authentication. The rebuild, run by a preparatory group under UPC registrar Alexander Ramsay and built on the EPO's infrastructure and interface conventions, set out to fix the practitioner experience.

Phase one, live from 8 July 2025, is deliberately narrow. To make the cutover, the existing CMS went dark from 3 July at 18:00 until 8 July, and on relaunch the new system initially handled opt-out and representative registrations only; infringement, revocation and the other substantive procedures migrate in later phases, with the legacy CMS kept running in parallel for everything not yet moved. The user-facing improvements are real and welcome:

• Software-based multifactor authentication replaces the physical smartcards and tokens that caused access delays (though smartcards persist for certifying and signing documents before upload).
• Flexible document uploads let parties file pleadings at any time, rather than within a rigid workflow window.
• A single-case workflow consolidates filings into one file, ending the parallel-case-number confusion of the old system.
• A familiar interface modelled on the EPO's existing online services, lowering the learning curve for representatives who already use EPO tools.

None of that is trivial — for the people who file at the UPC every week, it is a meaningful upgrade. But notice what the list is about: ingestion, authentication, case administration. It is a better front door for litigants, not a new window for the public.

What stayed the same for public data access

The UPC's transparency model is set by Rule 262 of the Rules of Procedure, and the new CMS did not touch it. That model has two tiers, and both survive the relaunch intact.

Decisions and orders are public. Under Rule 262.1(a), the Court's decisions and orders are made available to the public — in practice, as anonymised PDF documents posted to the register, with personal data redacted. This is the genuinely valuable part of UPC transparency, and it is more than several national systems offer. But it is a document drop, not a dataset: each ruling is an individual PDF, and the register offers no way to search across the full text of those decisions, filter them by outcome, or pull them in bulk.

Written pleadings are available only on a reasoned request. Under Rule 262.1(b), the parties' submissions and evidence are not posted; a member of the public must apply, explaining (in the Court's own framing) "the purpose of the request and why access is necessary." How that plays out has been uneven and slow. Commentators have documented stretches of many months before pleadings are released, and divisions have diverged sharply — the Munich section of the Central Division has refused access in cases where the Nordic-Baltic Regional Division has been markedly more liberal. For a researcher trying to assemble a consistent corpus, "on reasoned request, eventually, depending on the division" is not a data pipeline.

So the public-access picture after 8 July looks much as it did before. The table below sets the two states side by side.

The one moving part for developers was the public API, and it moved in the direction of less. From 8 July, three search endpoints were discontinued — the list of representatives, the list of representative entitlements, and the opt-out linked to a given patent. Three others were retained for the transitional period — search case, search case types, and list of languages — and, tellingly, those continued to retrieve data from the previous CMS rather than the new one. Even at their fullest, these endpoints returned case metadata: parties, case type, division, language. They never returned the thing that matters most for strategy — the result.

The gap the CMS leaves: it tells you a case exists, not who won

Here is the strategic point, stated plainly and without any criticism of the Court: a register is not an analytics platform, and the UPC's was never designed to be one. The new CMS is a tool for running litigation, and it does that better now than in 2023. But three things a litigant needs to make decisions remain entirely absent from the public system.

You cannot search the substance. With decisions published only as individual anonymised PDFs and no full-text index, there is no native way to ask "show me every Munich provisional-measures decision where the injunction was granted," or "every revocation counterclaim that ended in the patent held valid." You can find that a decision exists. You cannot query what it held.

You cannot see outcomes in aggregate. The register records that a case was filed, in which division, between which parties — but does not classify the result. There is no public field for infringed / not infringed, granted / denied, revoked / maintained, and so no public grant rate, revocation rate, or win record by party, firm or division. The official UPC framing — that, two years in, just over half of infringement actions decided on the merits ended with the patent valid and infringed, about a quarter not infringed, and about a quarter invalid — comes from the Court reading its own files, not from anything a third party can pull from the register.

You cannot get the data in bulk. There is no export, and after July the open API does less than before. Anyone who wants the structured picture has to build it — read each PDF, identify the parties and the patent, determine the disposition, store it in a queryable form.

That is, more or less, what our platform does: take the public record the CMS exposes — anonymised decisions, case metadata — and add the layer the register omits, a classified outcome for every decision plus normalised parties, representatives, division, action-type and patent context. As of our current snapshot that is 2,830 case records, of which 1,549 carry a classified substantive outcome, including 741 UPC decisions issued in 2025 that we have read and coded. We say this as a matter of fact, not complaint. The UPC sensibly prioritised a system that lets parties litigate cleanly; broad public analytics was never its remit, and the anonymised-decision register it provides is a genuine transparency win by the standards of European patent courts. Structured outcome data is simply a different job — turning a public record into something you can run a strategy against.

What this means for practitioners and data users

A few practical reads from the July changeover.

Treat the register as a primary source, not a research tool. The CMS is where you confirm a filing, a party, a division, a language — not where you benchmark a forum or scope an opponent. For that, the PDFs have to be read and classified, by you or by a tool that already has.

Mind the API if you automated anything. If your workflow touched the representative-list, representative-entitlements, or opt-out-by-patent endpoints, those stopped returning data on 8 July. The case-search endpoints still work for now, but run off the legacy CMS during the transition — build for the possibility that the underlying source moves.

Don't expect the analytics gap to close on its own. Nothing in the published rollout plan adds outcome classification, full-text search, or a bulk feed. Rule 262 governs access to documents, not analysis of them — so base rates like preliminary injunction grant rates by division, revocation success and settlement frequency will keep coming from a layer built on top of the register, not from the register itself.

What to watch

The thread most worth tracking through the rest of the rollout is the API. Phase one retired three endpoints and left three running off the old CMS as a stopgap. The open questions: when the substantive procedures finish migrating, will the retained search endpoints be re-pointed at the new system, quietly retired, or expanded? Will the EPO-built backend — the EPO runs some of the most capable patent-data infrastructure in the world — eventually expose richer, more reliable case search than the legacy API ever did? And does any of it ever cross the line from metadata (a case exists) into outcome (a case was won)? Our expectation, on the current plan, is that it will not — that the register will keep telling you a decision exists, and the work of telling you who won will keep living a layer above it. We will update this note if the residual endpoints, or a future phase, prove us wrong.

For now, the verdict on the new UPC CMS is straightforward: a better system for making a case, with the same gap for reading the docket. The front door got an upgrade. The data window did not.

Want the outcome layer the register doesn't expose — every UPC decision read, classified and queryable by division, party and result? That's exactly what our platform builds on top of the public record. Explore the dataset to benchmark your next filing.

Related reading

• How the Unified Patent Court Works: A Litigator's Guide to Divisions, Timelines & Outcomes
• The UPC at the Half-Year: Filing & Outcome Trends Through H1 2025
• Meet the Judges of the UPC: Panel Composition & Ruling Tendencies

Sources

• Unified Patent Court — "UPC Case Management System" (launch and rollout): https://www.unifiedpatentcourt.org/en/news/upc-case-management-system
• Unified Patent Court — "Update on public APIs following launch of first phase of the new CMS roll-out": https://www.unifiedpatentcourt.org/en/news/update-public-apis-following-launch-first-phase-new-cms-roll-out
• Unified Patent Court — "Progress in the development of new UPC Case Management System in cooperation with the EPO": https://www.unifiedpatentcourt.org/en/news/progress-development-new-upc-case-management-system-cooperation-epo
• Crowell & Moring — "The UPC's New Case Management System — What to Know": https://www.crowell.com/en/insights/client-alerts/the-upcs-new-case-management-systemwhat-to-know
• Hoffmann Eitle (reproducing UPC news) — "UPC launches first phase of new Case Management System on 8 July 2025": https://upc.hoffmanneitle.com/en/news/2025-07-01-upc-cms-first-phase
• Unified Patent Court — Rules of Procedure, Rule 262 (public access to the register): https://www.unifiedpatentcourt.org/en/court/rules-procedure
• UPClytics.com — internal dataset of classified UPC decisions (snapshot June 2026 build).