Bible study programs use a plethora of markup formats.
CrossWire Bible Society respects copyright. As such, conversion of material that is under copyright without permission from the copyright holders is not supported by The SWORD Project.
This page lists some of the more common file formats relevant to The SWORD Project, associated utilities, and other CrossWire projects.
- 1 SWORD modules
- 2 SWORD Input formats
- 3 The SWORD Project Utilities
- 4 Formats for which CrossWire maintains converters
- 5 Other Utilities
- 6 Optical Character Recognition
- 7 See also
Other than the source code for the SWORD API, there is no documentation for the file format of a SWORD module. There is no other documentation. The intention is that the SWORD API (or the JSword implementation) is used directly or via other language bindings.
Whilst our software is open source, our module file format is proprietary, in the sense that we see no need to document it and certainly no need to stick to it. We change it when we need to. We therefore do not encourage direct interaction with it, but firmly recommend use of the API (either C++ or Java). This is the place where we seek stability and consistency.
SWORD Input formats
The SWORD Project supports the following markup: OSIS, TEI, ThML, GBF and plain text.
Open Scripture Information Standard
The Open Scripture Information Standard (OSIS) is "a common format for many visions." It is an XML format for marking up scripture and related text, part of an initiative composed of translators, publishers, scholars, software manufacturers, and technical experts, coordinated by the Bible Technologies Group. It is co-sponsored by the American Bible Society and the Society of Biblical Literature.
In CrossWire, we can now validate OSIS files using our own custom adaptation of the OSIS schema. See OSIS 211 CR for details.
This markup format is recommended by the CrossWire Bible Society and can be used for creating all types of resources for The SWORD Project. Support for OSIS is actively maintained and support for any unsupported elements or features needed for a module you may be working on may be requested.
Prince XML is a proprietary software program that converts XML and HTML documents into PDF files by applying Cascading Style Sheets (CSS). It is developed by YesLogic, a small company based in Melbourne, Australia. It can be used to create high quality PDF Bibles from OSIS files. A paper by Jim Albright of Wycliffe Bible Translators was presented at BibleTech 2010 on using the open-source GUI companion for Prince XML, called Princess.
Text Encoding Initiative
TEI is an text encoding standard used broadly in the humanities. It is developed by the TEI Consortium.
A manual for the latest edition of TEI P5 can be found at P5: Guidelines for Electronic Text Encoding and Interchange.
CrossWire supports use of TEI for the encoding of dictionaries, encyclopedias, and similar works, as well as for free form books (GenBooks). We have developed our own version of the TEI P5 schema, available at http://www.crosswire.org/osis/teiP5osis.latest.xsd (http://www.crosswire.org/osis/teiP5osis.2.5.0.xsd is the latest at the time of writing). CrossWire's additions to the standard schema include the ability to create osisRefs and osisIDs to facilitate interoperability with OSIS documents and Bible reference tagging.
Further information can be found at the TEI Dictionaries page.
This plain-text format is used for by SWORD for import of Bibles. It consists of one verse per line, with an optional verse reference at the beginning. The vpl2mod utility may be used for import. VPL is deprecated in favor of the IMP format, which is more widely useful. The mod2vpl utility may be used for export to VPL. There is a command line switch to prepend the verse reference to each line.
This proprietary file format is used by SWORD for import of all types of modules. The three utilities imp2vs (for Bibles and verse-indexed commentaries), imp2ld (for lexicons, dictionaries, and daily-devotionals), and imp2gbs (for all other types of books) can be used to import IMP files to SWORD's native formats.
An IMP file consists of any number of entries. Each entry consists of a key line and any number of content lines. The key line consists of a line beginning with "$$$". For example, "$$$Gen 1:1" would be the key line for the Genesis 1:1 entry of a Bible or commentary module.
The content lines of an entry may consist of any text (provided that the first three characters of the line are not "$$$"). The internal markup of the content may be in any format supported by SWORD, namely OSIS for any module type or ThML for freeform books from CCEL.
With multiline content per key, any content line (except the last for each key) that would otherwise end with a space should be terminated by the continuation code \ (reverse solidus) to prevent joining words.
See also DevTools:Modules#IMP_Format.
- IMPORTANT: Please read http://www.crosswire.org/tracker/browse/MODTOOLS-92
CrossWire has a Perl script called imp2osis.pl, which will convert IMP to OSIS fairly well (except a few 'corner cases'). Whenever CrossWire receives an IMP submission, this is the first thing that is run, allowing CrossWire to do validation and other OSIS sanity checks. Some editing is usually necessary after converting an IMP file to an OSIS XML file. For example, the attribute canonical is omitted from the osisText and all <div> elements, and the language attribute xml:lang defaults to "en".
Theological Markup Language (deprecated)
This format is a variant of XML based on TEI and ThML, developed by and for the Christian Classics Ethereal Library. The specifications for this markup format are available at http://www.ccel.org/ThML/.
This markup format is used in some SWORD resources, but only the creation of free-form "General book" modules based on existing CCEL resources is currently supported. Other works and new works should be created using the OSIS or TEI format.
General Bible Format (deprecated)
This markup format is intended as an aid to preparing Bible texts (specifically the WEB and WEB:ME) for use with various Bible search programs. The complete specification is at http://www.ebible.org/bible/gbf.htm.
This markup format was previously used for some SWORD modules but is now deprecated in favor of OSIS. The rudimentary gbf2osis.pl Perl utility may be used to convert GBF to OSIS for import to SWORD's native format. Adyeth hosts a gbf2osis Python utility that he wrote to convert the GBF texts from ebible.org to OSIS. See .
The SWORD Project Utilities
Module Creation Tools
- imp2gbs - imports free-form General books in IMP format to SWORD format †
- imp2ld - imports lexicons, dictionaries, and daily devotionals in IMP format to SWORD format †
- imp2vs - imports Bibles and commentaries in IMP format to SWORD format †
- vpl2mod - imports Bibles and commentaries in Verse-Per-Line format to SWORD format †
- osis2mod - imports Bibles and commentaries in OSIS format to SWORD format †
- xml2gbs - imports free-form General books in OSIS or ThML format to SWORD format †
- mod2imp - creates an IMP file from an installed module †
- emptyvss - exports a list of verses missing from the module (useful for testing modules during development) †
- gbf2osis.pl - a PERL utility for converting GBF to OSIS †
- step2vpl - export a STEP book in Verse-Per-Line (VPL) format †
- thml2osis - converts ThML to OSIS format.
- mod2osis - creates an OSIS file from an installed module †
- vs2osisref - returns the osisRef of a given (text form) verse reference †
- xml2gbs - imports free-form General books in OSIS or ThML format to SWORD format †
- cipherraw - used to encipher SWORD modules †
- diatheke - a basic CLI SWORD front-end †
- mkfastmod - creates a search index for a module †
- mod2zmod - creates a compressed module from an installed module †
- mod2vpl - exports the module to VPL format †
- modwrite - outputs the module contents in VPL format †
- treeidxutil - needs a description †
- genbookutil - needs a description †
Notes on SWORD Tools
- If you have Xiphos installed in Windows, the Sword utilities are available in the Xiphos\bin folder.
- The latest binaries may be found here, though currently without cipherraw.exe (Link broken!)
- EOLs should be either Unix style (LF) or Windows style (CRLF). Text files with Mac style EOLs (CR) may give rise to errors or other unexpected behaviour.
- The IMP file may contain a residue of XML markup
- Aside: To create a list of installed modules with descriptions, enter the following command, optionally redirecting stderr to a log file.
mkfastmod /? 2>mkfastmod.log
- The VPL file may contain a residue of XML markup
Recommended Non-SWORD Utilities
- uconv - a utility from ICU for converting between various character encodings, perform normalization, transliterate texts, etc. (It's similar to iconv, but much, much more powerful.) uconv.exe is part of the sword utilities
- xmllint - a utility (part of the libxml2 distribution) for validating XML documents *
Formats for which CrossWire maintains converters
The SWORD Project uses primary source e-texts. These texts come in numerous formats. CrossWire maintains converters for a number of formats, described below. The converters may target other markup formats, e.g. TEI or OSIS, or may simply export binary data to text, as is the case with our STEP exporter. Specific discussion of each of the available converters is found elsewhere on this page.
This plain-text format is a common internal-use format within Bible translation agencies and Bible societies. It is the native format of ParaTExt. Paratext is used by more than 60% of all Bible translators world-wide. The current release is Paratext 7.5.
USFM uses a separate file for each Bible book. USFM is also supported by the open-source program called Bibledit. There are examples of Bibles in USFM format available for download at . These include the KJV, ASV, and WEB Bibles.
USFM is one of the formats that can be used by Go Bible Creator.
- This replaces our earlier Perl script usfm2osis.pl.
Following an agreement made in July 2008 with the program's author Jolon Faichney, Go Bible was adopted by CrossWire as its Java ME software project.
To achieve the navigation speed and general ease of use on even the simplest of Java mobile phones, Go Bible data is fully indexed, as well as being compressed (as are all JAR files). The format is described in Go Bible data format. Go Bible data is structured as Book | Chapter | Verse text and does not support notes, headings and cross-references, etc. The developer kit Go Bible Creator can take either USFM, ThML or OSIS as the source text format, but they usually have to be made specially suitable. For example, OSIS files produced by Snowfall Software's SFMToOSIS script are not structured the same. Work has begun to make an XSLT script to convert such OSIS XML files to the format suitable for Go Bible. Go Bible Creator version 2.3.2 and onwards can take a folder of USFM files as the source text format.
Go Bible source code is now available here on the CrossWire Repository. To access this you will need to have an account.
GoBibleDataFormat is being extended in the SymScroll branch.
These are not part of The SWORD Project, but may be useful. A link is given for each.
Go Bible utilities
- Go Bible Creator - a Java SE program for converting either ThML or OSIS or USFM to Go Bible. It is being enhanced by SIL to be capable of converting source text in XHTML-TE format.
- Go Bible Creator USFM Preprocessor – This is a tool to parse through and identify, correct and publish USFM file formats into a file format that can easily be put into the Go Bible mobile phone program.
- CCEL Desktop - a program for viewing and developing CCEL books 
Optical Character Recognition
Text development activities may be greatly assisted by using OCR software. This section will list OCR programs that CrossWire volunteers have found useful. Proprietary programs should not be listed here, the preference at CrossWire being to use free and/or open-source software.
- Tesseract is a free optical character recognition engine. It was originally developed at Hewlett-Packard from 1985 until 1995. After ten years with no development, Hewlett Packard and UNLV released it in 2005. Tesseract is currently developed by Google and released under the Apache License, Version 2.0. Tesseract is considered one of the the most accurate free software OCR engines currently available.
- VietOCR – A Java/.NET GUI frontend for Tesseract OCR engine. Supports optical character recognition for Vietnamese language. Tip: Visit  to read how Moheb Mekhaiel adapted VietOCR to scan Coptic documents.