User Rating 0.0 ★★★★★

Total Usage 0 times

Category Document Converters

Drop .po files here or click to browse Supports multiple files

AMD module wrapper Use msgctxt as key Exclude fuzzy entries Include header metadata Merge files into single JSON

Indentation

Is this tool helpful?

Your feedback helps us improve.

★ ★ ★ ★ ★

About

GNU gettext .po files remain the dominant interchange format for software localization, yet modern front-end stacks consume JSON. A manual conversion introduces encoding errors, drops plural forms, and silently mishandles multiline msgstr continuations. This tool parses the full PO grammar - including msgctxt disambiguation, msgstr[N] plural indices, and escaped sequences (\n, \t, \") - then emits clean JSON or an AMD-wrapped JavaScript module. All parsing runs locally in your browser. No file is uploaded to any server.

Limitations: the converter assumes UTF-8 encoding. Binary .mo files are not supported. Obsolete entries (prefixed with #~) are intentionally excluded from output. Fuzzy-flagged entries (marked #, fuzzy) are included by default but can be excluded via the option toggle. Pro tip: if your PO file uses msgctxt for key namespacing, enable the "Use msgctxt as key" option to avoid collisions between identical msgid values in different contexts.

Formulas

The converter operates as a line-by-line state machine. Each line is classified by its prefix and the parser transitions between states accordingly.

parse(line) →

{

COMMENT if line starts with #MSGCTXT if line starts with msgctxtMSGID if line starts with msgidMSGSTR if line starts with msgstrCONTINUATION if line starts with "

Continuation lines (lines beginning with a double quote) are appended to whichever directive was active in the previous state. The final JSON key is determined by the mode:

key =

{

msgctxt if useMsgctxtAsKey = TRUE ∧ msgctxt ≠ NULLmsgid otherwise

For plural entries, the value becomes an array:

value =

{

string if single msgstr[msgstr[0], msgstr[1], …, msgstr[N]] if plural

Where key is the lookup identifier and value is the translated content. The header entry (empty msgid) is always excluded from output. When AMD wrapping is enabled, the output becomes: define(function() { return JSON ; });

Reference Data

PO Directive	Purpose	JSON Mapping	Notes
msgid	Source string (key)	Object key (default mode)	Empty msgid = file header
msgstr	Translated string	Object value (string)	Empty msgstr means untranslated
msgctxt	Disambiguation context	Used as key when option enabled	Avoids duplicate msgid collisions
msgid_plural	Plural source string	Ignored (informational)	Plural source not needed in output
msgstr[0]	Singular translation	Array index 0	Required when plurals present
msgstr[N]	Nth plural form	Array index N	Count depends on language
# (comment)	Translator comment	Not included in output	Informational only
#.	Extracted comment	Not included in output	Auto-generated by xgettext
#:	Source reference	Not included in output	File:line location
#,	Flags (e.g., fuzzy)	Entry skipped if fuzzy excluded	Comma-separated flags
#~	Obsolete entry	Always excluded	Deprecated translations
#\|	Previous msgid	Not included in output	Used by tools for diffing
\n	Newline escape	Literal newline in value	Common in multiline strings
\t	Tab escape	Literal tab in value	Rare but valid
\"	Escaped quote	Literal " in value	Required inside quoted strings
\\	Escaped backslash	Literal \ in value	Standard C escape
Plural-Forms	Header: plural rule	Not included (header entry)	e.g., nplurals=3; plural=...
Content-Type	Header: charset	Not included (header entry)	Should be text/plain; charset=UTF-8
Language	Header: language code	Not included (header entry)	ISO 639-1 code (e.g., fr)

Frequently Asked Questions

PO files allow a directive to span multiple lines. The first line contains the directive keyword followed by an empty string (e.g., msgstr ""), and subsequent lines contain only quoted strings (e.g., "continued text"). The parser detects continuation lines - any line starting with a double quote character while not beginning with a keyword - and appends their unquoted content to the current active directive. Escaped newlines (\n) inside the quoted text are converted to actual newline characters in the JSON output.

In default mode (using msgid as key), the second entry overwrites the first because JSON object keys must be unique. This is a known limitation of flat key-value mapping. Enable the "Use msgctxt as key" option to use the disambiguation context as the JSON key instead, which preserves both entries under their distinct msgctxt values. If msgctxt is absent for an entry in this mode, the converter falls back to msgid.

By default, yes. Fuzzy entries (marked with #, fuzzy in comments) represent translations that need human review but may still be partially correct. The converter provides a toggle to exclude fuzzy entries entirely. When excluded, any entry whose flags contain the word "fuzzy" is omitted from the JSON output. This is useful for production builds where only verified translations should ship.

When a PO entry contains msgid_plural and multiple msgstr[N] directives, the converter produces an array value instead of a string. The array indices correspond to the plural form indices: msgstr[0] becomes index 0, msgstr[1] becomes index 1, and so on. The number of plural forms depends on the target language (e.g., 2 for English, 3 for Polish, 6 for Arabic). The Plural-Forms header directive defines the rule but is not included in the output.

Yes. When you upload multiple PO files, the converter can produce either separate JSON outputs (one per file) or a single merged JSON object where each file's translations are nested under a namespace key derived from the filename. For example, uploading messages.po and errors.po produces {"messages": {...}, "errors": {...}}. This mirrors the grunt-po-json behavior of mapping multiple sources to named properties.

No. The header entry (the entry with an empty msgid) contains metadata like Project-Id-Version, Content-Type, and Plural-Forms. This metadata is parsed internally to detect encoding and plural rules, but it is excluded from the JSON output. The header is not a translation entry and would pollute the key-value map with an empty-string key. If you need the header, enable the "Include header" option which adds it under a __header__ key.