Line4 logoLine4

Search line4.ai

What is an XBRL taxonomy package?

An XBRL taxonomy package is a zip file that carries a taxonomy with a standard label: one top-level folder, and inside it a META-INF/taxonomyPackage.xml that names the package and lists its entry points, the files a filing starts from. An optional META-INF/catalog.xml maps the taxonomy’s web addresses to files in the zip, so software can load it offline. XBRL International’s Taxonomy Packages 1.0 (April 2016) defines it. Outside Europe, the FFIEC offers each Call Report taxonomy as a compressed download, and Australia’s APRA publishes zip files holding an Excel taxonomy and XSD files.

Guide · 5 min read · Updated 21 November 2026 · By Jess Luper

What a taxonomy package is

A taxonomy is hundreds or thousands of files, and they point at each other by web address. Before 2016 every publisher shipped its own zip, and software had to be told which file to open first and where the web addresses really lived.

Taxonomy Packages 1.0 fixes that with a small manifest. In XBRL International’s words it lets “compliant tools … identify the entry points automatically”, and treat the package “as an offline copy of taxonomies published at an Internet location, without the need for additional configuration”.

01One top-level folderThe zip MUST contain a single top-level directory, with everything else inside it (section 3.1). Naming it after the zip is recommended, not required.
02META-INF/taxonomyPackage.xmlThe package MUST have this file (section 3.1). Only its identifier, a URI for the package, is required; name, version, publisher, publication date and entry points are optional.
03Entry pointsA package MAY list entry points: named starting files, each a taxonomy schema or linkbase, from which software discovers the rest of the taxonomy (section 3.2.2).
04META-INF/catalog.xmlOptional. Its rewriteURI lines map a published web address, such as http://www.eba.europa.eu/eu/fr/xbrl/, to a folder inside the zip (section 3.3).
2016Taxonomy Packages 1.0 became an XBRL International Recommendation (19 April)XBRL International →
1file every package must have: META-INF/taxonomyPackage.xml, with an identifierTaxonomy Packages 1.0, section 3.2 →
62EBA module rule sets served by Line4 Check, each tied to one module entry pointLine4 count of the Line4 Check bundles, November 2026

An example: how a filing finds its taxonomy

A filing never carries the taxonomy. It names an entry point by its published web address, and the supervisor’s software resolves that address, ideally from a package it already holds.

Two real entry points, from public sample filings
EBA COREP leverage ratio (xBRL-CSV)
report.json extends http://www.eba.europa.eu/eu/fr/xbrl/crr/fws/corep/4.2/mod/corep_lr.json, the module’s JSON entry point in framework 4.2
EIOPA Solvency II annual return (XBRL-XML)
schemaRef points to http://eiopa.europa.eu/eu/xbrl/s2md/fws/solvency/solvency2/2026-06-30/mod/arb.xsd, a schema entry point in taxonomy 2.10.0
What the catalog does
Maps the http:// start of each address to a folder in the package, so the file is read from disk, not the internet
What the entry point decides
Which templates, concepts and validation rules apply to this filing, and nothing else
What a supervisor’s system learns from it

Every EBA address in this filing is answered from the package on disk: same files every time, no network call, no surprise if a web page moves.

A rewriteURI line in catalog.xml (illustration)Show<rewriteURI uriStartString="http://www.eba.europa.eu/eu/fr/xbrl/" rewritePrefix="../www.eba.europa.eu/eu/fr/xbrl/"/>

Entry point addresses from the EBA and EIOPA public samples served by Line4 Check. The catalog line is a Line4 illustration of the pattern section 3.3 describes, not copied from a published package.

Entry points: one per return

A supervisory taxonomy is really many taxonomies sharing one dictionary. Each return, such as COREP own funds or a Solvency II quarterly, has its own entry point, and a filing must name exactly the one it follows.

That makes the entry point the most important line in a filing. Name the wrong version and the filing is checked against the wrong rules. Line4 Check reads it first and picks the rule set built for that module and version.

A filing names one entry point. Everything else follows from the package behind it.

What it means for supervisors

If you publish a taxonomy, publish it as a Taxonomy Packages 1.0 zip, list every entry point with a name and version, and include a catalog. Firms and vendors can then load it in any compliant tool with no setup, and you stop answering “which file do I open?” questions.

If you receive filings, load taxonomies only from packages you hold, never from the live web. The check then depends on a fixed, versioned copy, so the same filing gets the same answer next year, and a filing that names an entry point you have not loaded is stopped on arrival with a plain message.

Where AI fits, and where people decide

AI can read a new taxonomy package the day it is published: list the entry points that changed, compare them with the last release and draft the change note for firms.

Which entry points and versions you accept is a supervisory decision, written in your filing rules. Loading and resolving the package is a fixed process, so the result is repeatable and kept with each filing.

Put it to workBefore your next taxonomy release
  1. 1Is your taxonomy published as a Taxonomy Packages 1.0 zip, with a taxonomyPackage.xml?
  2. 2Does it list every entry point, with a name and a version?
  3. 3Does it include a catalog.xml, so tools can load it offline?
  4. 4Does your collection platform load taxonomies only from packages you hold?
  5. 5Do filings that name an unknown entry point get a plain message on arrival?

Where Line4 fits

Line4 is SupTech software for financial regulators. It collects returns in XBRL, Excel and CSV by portal, API or MCP, validates them against your taxonomies and rules at the point of submission (see XBRL software for regulators), and puts AI to work reading them, with supervisors making every decision. If you work with a standard we have not yet run in production, we will say so plainly and show you how we would approach it.

Questions

What is an XBRL taxonomy package?

A zip file that carries an XBRL taxonomy with a standard manifest, META-INF/taxonomyPackage.xml, which identifies the package and can list its entry points. An optional META-INF/catalog.xml maps the taxonomy’s web addresses to files inside the zip. It is defined by XBRL International’s Taxonomy Packages 1.0, a Recommendation since 19 April 2016.

What is an entry point in XBRL?

The file a filing starts from, so software knows which part of a taxonomy, and so which templates and rules, apply. Under Taxonomy Packages 1.0 an entry point is one or more entryPointDocument links, each a taxonomy schema or linkbase. For EBA returns it is a JSON file for xBRL-CSV or a schema for XBRL-XML, one per reporting module.

What does catalog.xml do?

It remaps web addresses. Each rewriteURI line says that addresses starting with, for example, http://www.eba.europa.eu/eu/fr/xbrl/ should be read from a folder inside the package. Processors SHOULD apply the remappings, so a compliant tool can load the taxonomy without going online. The file is optional.

Which parts of taxonomyPackage.xml are required?

Only the identifier, a URI that uniquely identifies the package. Name, description, version, licence, publisher, publication date, entry points, superseded packages and versioning reports are all optional in the specification, though a regulator publishing a taxonomy should fill them in.

What is the difference between a taxonomy package and a report package?

A taxonomy package carries the dictionary a filing is checked against; a report package carries the filing itself. A report package may also be a valid taxonomy package, which is how an Inline XBRL report can travel with its own extension taxonomy.

Can I load a taxonomy without a package?

Yes, by fetching every file from its published web address, but it is slow, needs a network, and depends on the publisher’s site staying the same. Supervisors should load from packages they hold, so each check uses a fixed, versioned copy.

Does Line4 Check use taxonomy packages?

Line4 builds its rule sets from the taxonomies the EBA, EIOPA and the Bank of England publish, one per module and version. Line4 Check reads the entry point in your filing and runs the matching rules in your browser, without uploading the file.

How do regulators outside Europe publish their taxonomies?

Not always as XBRL taxonomy packages. The FFIEC offers each Call Report taxonomy as a compressed download, one for every quarter back to March 2001. Australia’s APRA publishes zip files holding an Excel reporting taxonomy and XSD files, which check the XML returns uploaded to APRA Connect. Line4 loads XBRL packages, Excel taxonomies and XSD schemas, so a regulator can bring whichever kind it has.

Share this guideShare on LinkedIn
Was this useful?

For supervisors

New guides and Insights, by email

One short email when we publish something worth your time, from Jess Luper. No spam, unsubscribe any time.