RIS File Documentation

Summary

An .ris file is a Research Information Systems citation file: a plain-text bibliographic record in which each field is a two-letter tag followed by two spaces, a hyphen and a space. Its MIME type is application/x-research-info-systems. It is the near-universal export format of Google Scholar, PubMed, publishers and library databases. You import it into a reference manager such as Zotero, Mendeley or EndNote rather than “opening” it; any text editor can display its raw tags.

Technical details

FeatureValue
Full nameResearch Information Systems citation file
File extension.ris
MIME typeapplication/x-research-info-systems
Format typeTagged plain-text bibliographic format
EncodingASCII / UTF-8, one tag per line
Tag syntaxTwo uppercase chars, two spaces, hyphen, space (XX  - )
Record startTY  - (type of reference)
Record endER  - (end of reference)
DeveloperResearch Information Systems (Reference Manager)
Now associated withClarivate / EndNote lineage
IntroducedEarly 1990s (Reference Manager)
Open standardPartial — documented tag set, no formal standards body
Magic numberNone (plain text; identified by the TY  - first line)
Multiple recordsYes — records concatenated, each TY…ER
Exported byGoogle Scholar, PubMed, Scopus, Web of Science, IEEE Xplore
Imported byZotero, Mendeley, EndNote, JabRef, Papers
Related extensions.bib, .enw, .enl, .nbib, .csl
Specificationen.wikipedia.org/wiki/RIS_(file_format)
Syntax at a glance

An RIS file is plain text with no binary signature. Every line is a tag line: two uppercase characters, then two spaces, a hyphen and a space (XX  - ), then the field value. A record always opens with TY  - (type of reference, e.g. TY  - JOUR) and always closes with ER  - on its own line. Repeatable tags such as AU  - (author) appear once per value on consecutive lines. A single file may hold one reference or thousands, simply concatenated. Open it in any text editor to read the raw tags before importing.

What is a RIS file?

RIS is a tagged plain-text format for bibliographic citations, created in the early 1990s by the company Research Information Systems for its Reference Manager software and named after that company. A .ris file describes references — journal articles, books, conference papers — not the documents themselves. Each reference is a block of tagged lines, and one file can carry a single citation or an entire database export of thousands. Its MIME type is application/x-research-info-systems.

The format was never ratified by a standards body, yet it became the lingua franca of academic citation exchange. Almost every scholarly source offers an RIS download — Google Scholar’s “Cite → RIS/EndNote”, ScienceDirect, Scopus, Web of Science, SpringerLink, IEEE Xplore, library catalogues — and every reference manager imports it. It endures for one reason: it is trivially simple, human-readable text, so any program can emit or parse it in a few lines of code. Reference Manager itself was discontinued at the end of 2016; its sibling EndNote, now owned by Clarivate, carries the lineage.

The tag line: two letters, two spaces, hyphen, space

The whole format rests on one line grammar. Every meaningful line is a tag line with a fixed six-character prefix: two uppercase letters or digits, then exactly two spaces, a hyphen, and one space, then the value.

T Y _ _ - _ J O U R
│ │ │ │ │ │ │
│ │ │ │ │ │ └─ value begins here
│ │ │ │ │ └─── one space
│ │ │ │ └───── hyphen
│ │ └─┴─────── two spaces
└─┴─────────── two-character tag
(_ marks a space character)

That spacing is exact and part of the format: the delimiter is literally space space hyphen space, so the value always starts at column 7. Parsers key on this fixed offset, which is why a stray reformatting that collapses the double space can break an import. Lines that do not match the tag pattern are continuations of the previous field’s value or are ignored. Tags are case-sensitive uppercase, and the same tag may repeat when a field has several values.

TY and ER: the record delimiters

Two tags are structural rather than descriptive. TY (type of reference) must be the first line of every record, and ER (end of reference) must be the last; its value is empty. Everything between them belongs to one citation, and the next TY begins the next.

TY  - JOUR
AU  - Shannon, Claude E.
TI  - A Mathematical Theory of Communication
JO  - Bell System Technical Journal
PY  - 1948
VL  - 27
SP  - 379
EP  - 423
DO  - 10.1002/j.1538-7305.1948.tb01338.x
ER  -

The value on the TY line is a reference-type code that tells the importing manager what kind of work this is and therefore which fields to expect. Common codes include JOUR (journal article), BOOK, CHAP (book chapter), CONF (conference paper), THES (thesis), RPRT (report), ELEC (web page) and GEN (generic). Because ER unambiguously closes a record, concatenating exports from different databases into one file works without any wrapper: the reader just loops from each TY to its matching ER.

The common data tags and repeatable fields

Between TY and ER sit the descriptive tags. Some are single-valued; others, such as authors and keywords, are repeatable — each value gets its own line with the same tag rather than being comma-joined, which keeps author names with commas in them unambiguous.

TagFieldNotes
AU / A1AuthorRepeatable; one per line, Last, First
TI / T1Title of the workPrimary title
T2 / JOContainer / journalJournal or book title
PY / Y1Publication year / dateYYYY or YYYY/MM/DD/
VL / ISVolume / issue
SP / EPStart / end page
DODOIDigital Object Identifier
SNISSN / ISBNSerial or book number
URURLLink to the resource
AB / KWAbstract / keywordKW repeatable, one per line

A quirk worth knowing: several fields have two accepted tags because different vendors settled on different ones. The title may arrive as TI or T1; the year as PY or Y1; the container as JO, T2 or the abbreviation JF. Good importers treat these as synonyms. Date values use a slash-separated YYYY/MM/DD/other pattern where trailing components can be omitted, so 1948/// and 1948 both mean the year alone.

Why you import a RIS file rather than open it

Double-clicking a .ris on Windows often does nothing, because the OS has no default application registered for the extension until a reference manager is installed. That is the source of most “my citation won’t open” confusion: the file is not broken, it simply has no handler. RIS is not meant to be read as a document; it is meant to be imported into a citation manager, which parses the tags into a structured library entry and can then format it in any citation style (APA, MLA, Chicago) for a paper.

Zotero (free, open-source) registers itself as the .ris handler so that double-clicking imports the references; Mendeley, EndNote and JabRef all read RIS through File → Import as well. To simply inspect what a file contains before importing, open it in any text editor — it is plain text. PubMed publishes a closely related tagged format called NBIB (MEDLINE) that uses different tags but the same idea of one tag per line.

RIS and BibTeX: two text citation formats

The most common conversion is RIS to BibTeX, because LaTeX users need .bib. Both store citations as plain text, but their grammar differs. RIS uses the two-letter tag-per-line model shown above; BibTeX uses typed entries with braced key–value pairs:

@article{shannon1948,
  author  = {Shannon, Claude E.},
  title   = {A Mathematical Theory of Communication},
  journal = {Bell System Technical Journal},
  year    = {1948}, volume = {27},
  pages   = {379--423}
}

The field mapping between them is well defined: RIS TY - JOUR becomes BibTeX @article, AU becomes author, PY becomes year. Importing a RIS into Zotero or JabRef and exporting as BibTeX is the reliable route; the command-line bibutils suite does the same with ris2xml | xml2bib. A few RIS tags have no direct BibTeX equivalent and are dropped or moved to a note field, which is the usual cause of minor metadata loss in the conversion.

References