Statement of Purpose:
Extracts data for multi-company comparison via desktop application, web service, or API. Provides fully confomant implementation of XBRL processor including validation, formula processing, versioining, and rendering. Both XBRL Formula and Sphinx 2.0 Rules are supported.Arelle’s validation includes XBRL 2.1, XBRL dimensions, SEC Edgar Filer Manual, IFRS Global Filer Manual, HMRC, and SBR. Arelle is in production use by government agencies, filers, banks, and individuals.Resources:
Source files URL
Screencast demo
Installable Application
License (Apache-2)
User’s Manual (English)
Manual de Usuario (Spanish)
Arelle Unified Model – published paper
GUI Operation
Command Line Operation
Web Services API

Herman Fischer

Instructions for Use

Desktop installation and step-through instructions are available in User’s Guides in English and in Spanish, and on the web site documentation pages installation and GUI operation.

Here is a short excerpt to load, explore, validate, or run test cases:

  1. Select open file () or open web () and choose the
    • XBRL instance or inline XBRL instance
    • XBRL DTS schema or linkbase
    • Archive file (zip, xfd, or frm). (When chosing an archive file, a secondary dialog will allow chosing the file to open within the archive.)
    • Test case file
    • Test suite index file
  2. You may browse contents of the XBRL instance, XBRL DTS, test case, or test suite, in the appropriate tab panes. Selecting entries in a tab pane will display any detailed properties in the Properties tab and also synchronize any linked entries in other tab panes.
  3. Select validate ().
  4. Messages in the messages pane may be saved to a file with the tools menu, save messages to file.
  5. close () the input

Here is a short excerpt to get ratios of multi-company submissions by the RSS feed and Sphinx Rules:

  1. Activate Sphinx Rules plugin: Help->Manage plug-ins, Find plug-in modules: Locally. Select examples\plugin\sphinx\
  2. Select RSS watch with Sphinx 2.0 multi-company ratios file: Tools->RSS Feed->Options, click folder and select sphinx rules file (link to example file). Check the validate sphinx checkbox.

Improves access

Investor stakeholders have complex needs for corporate data (and by utilizing XBRL they can access the data in real time as it is published by the SEC, without waiting minutes or even days for analysts at the large aggregators to post information). The real-time stakeholder has integrated a sophisticated workflow of small products to aid in decision making, which can now access information in as timely a manner as the large institutions.

XBRL various tremendously in how it provides access to corporate data, from its use in US SEC filings, where there’s a great freedom in mapping data points to standardized taxonomy concepts, to extremily high volumes of filings in other environments which might be more constrained (such as UK HMRC, Eurofiling, and EDInet).

Primary Arelle users are currently primarily the producers of filings, and the governmental agencies and banks who receive and validate them. Stakeholders now have access to the same tooling.

The RSS watch feature allows casual and serious users to be alerted for text patterns in filings and the results of processing Sphinx reporting rules against filings, in near-real time as they are published by the US SEC.


Arelle is a complete XBRL processor including formula, rendering, and versioning. It provides these facilities three ways, to desktop users and local tools integrated on the desktop (such as Excel and Quickbooks), and it provides server implementation of validation, formula processing, and web support.


This Challenge submission provides extraction of derived data and ratios from individual SEC submissions, which can then be incorporated in the workflow of analysts and user’s existing tool flow to be combined with extractions from other submissions..

This extraction is highlighted here using Sphinx, which processes a single instance at a time, and can extract data from multiple reported time periods and dimensional contexts (such as business units) in a submission, and combined externally with those of other submissions.

Sphinx allows simple analytics to be expressed in business terms, such as by coding a ratio expressed in terms of the reported fact names:

report CurrentRatio
AssetsCurrent[] / LiabilitiesCurrent[]

The processor reports the values aligned up by period and dimensions. Sample ratios are provided in the link under (2) above. More sophistication is possible, such as to specify alternative choices of names under which items might be reported, and dimensional and period qualification. Some business issues are discussed by Paul Warren in his blog on Sphinx.

The value of Sphinx is the expressive notation is in business reporting terms, and that can accommodate the nature of investment decision information required.


Arelle is an original XBRL product completely implemented in Python, the language preferred for financial modeling and analysis. Its architecture provides the first XBRL processor based on a unified model that integrates all aspects of XBRL, from instances (traditional and inline), their DTSes and formulas, to test suites (treated as first class model objects), RSS feeds, and versioning reports. Its architecture is living, being adopted as XBRL migrates from a producer XML syntax-centric paradigm, to the AMTF data points and OMG CWM consumer supportive paradigm.

Arelle is also original in uniquely supporting multiple modes of operation, from desktop users by graphical user interface, to integrated workflow managed systems accessing XBRL by RESTful web services. Arelle supports users both large and small, including features that allow individual Excel and Quickbooks users to easily integrate their products, and integration to languages and environments (such as Visual Basic, C#, and Java).


Arelle now supports a plug-in architecture so that supplemental functionality can be incrementally added without modifying the core source code. All portions are open source.

For the individual CPA practitioner, bloggers are active with various forms of integrating Excel and Arelle, some using web services and others command line intefaces. The QuickBooks interface supports practitioners with GlobalLedger XBRL of trial balance, journal, and ledger, which can be converted to financial reports.

For the reviewer of filings, Arelle supports early-phase consumption, of individual instances by the RSS watch and selected instance loading. The Unified Model paper (in links box) explores data mining in this manner.

For the user of data points methodology and table rendering, Arelle is the premeir implementation of the XBRL Table Linkbase and rendering engine. Arelle has been selected by the European Banking Authority (EBA) as the open source processor for 2013 FINREP.

The XBRL Formula Linkbase, which was available within last year’s challenge entry of this project, additionally supports multi-instance Formula Processing, and can be given formulas to receive data from multiple separately independent SEC submissions. These formulas are more complex to create than the Sphinx approach, and are mostly used in the European, Australasian, and Japanese environments. (The advice to this project has not been enthusiastic about having North American investors and analysts writing XBRL Formula Linkbases, though the same advice is greatly enthusiastic about the same usage by Sphinx expressions).

A thorough integration of multiple instances and multiple reporting periods of separate submissions is intended using a derivation of the XBRL Abstract Model, of which this submitter is an author. The Abstract Model is based on OMG’s Computer Warehouse Metamodel (CWM) which includes OLAP as a subset. Attempts by this contest submission of Arelle to utilize existing OLAP implementations have not proved useful for SEC filings because such filings require dynamic dimension creation (each submission has a different taxonomy), a feature not really achievable in commercial OLAP implementations. Arelle is now pursuing non-SQL databases (based on highly efficient large-data graph models, such as Hadoop). This is a long multi-year effort significantly outside of the scope of this contest submission.

Arelle pioneers early consumption features, and its contributors are very active in the evolution of XBRL from filing production to large data set consumption.

The future of consumption is clear from the XBRL conference technical tracks and investment in technology to conform and extract, transform, and load XBRL filings into business warehousing technology where traditional business analytics can process large volumes of filings. As this technology progresses in the near term, the members of the Arelle community, both contributor and consumer, intend to be at the forefront.