See exactly which word matches which across stacked lines. Add rows for glosses or IPA, click
a word then its match on the line above or below, and export or share the diagram. Good for
lessons, posts, or conlang notes.
What this tool does
Word Aligner is a free word-by-word translation visualizer: it shows which words in one line correspond to which
words on the next. Type or paste text in the line editor, add more rows
with + Add line when you want glosses, IPA, or another tier, then put lines in the order you need. In the preview, click a word, then
click a match on the line directly above or below;
only those adjacent rows link, so reorder lines with the arrows if something is out of reach.
Linguists call this a word alignment or bitext alignment; most people just want to see which
words match.
Guides
Why is a visual word alignment clearer than plain text?
A plain word-by-word translation under each line breaks down the moment the translation changes
word order, uses one word where the source uses three, or collapses several source words into a
single morpheme. Curved connectors solve all three cases: lines can cross, fan out from one word
to several, or converge from several words onto one. That is why a bilingual sentence alignment
is easier to read than an interlinear row when you care about structure, not only meaning.
How do learners and teachers use word alignment?
Learners see at a glance why a translation says what it does: which source word became which
target word, which words are dropped, and where the target language moves things around.
Teachers can build handouts, slides, or flashcards by exporting PNG or SVG and embed the
finished diagram into a lesson without retyping anything.
How do conlangers use it for glosses?
For conlangers, the visualizer is a lightweight way to show how a constructed language maps onto
English in a specific example. Add extra lines for glosses or IPA (and stack
them next to the sentences they annotate), tune how text splits into word-sized boxes under Settings → Tokens, and export
the result for a blog post, a forum thread, or a conlang community share. The conlang glossing guide walks through
custom scripts, an IPA tier, and exports.
What is the difference between word alignment and interlinear translation?
Interlinear translators place a translation directly under each source word. That display is
compact and good for reading, but it hides reordering: if the translation swaps word order, the
row underneath misrepresents which word corresponds to which. Word alignment keeps both
sentences on their own line and draws connectors between them, so reorderings, splits, and
merges are obvious.
What is the difference between word alignment and parallel text?
Parallel text usually means side-by-side bilingual reading: two columns, or a source paragraph
next to a translated paragraph, meant for studying in long form. Word Aligner does a different
job, one sentence pair at a time, with explicit connectors showing which token corresponds to
which. Parallel text shows that two passages translate each other; word alignment shows exactly
which word answers to which, including splits, merges, and reordering.
Glossing notation at a glance
The labels you meet most often are NOM (subject), ACC (object), GEN (possessor), PST (past), PFV (completed
action), 1SG (first-person singular), PL (plural), and NEG (negation).
A few marks line the gloss up with the source: a hyphen (-)
separates morphemes, a period (.) joins one form that carries
several meanings (come.out), a tilde (~) marks reduplication, an overt ø marks a morpheme with meaning
but no sound, and a backslash (\) marks a change inside the word.
The full set is on the glossing abbreviations cheat sheet.
Glossing guides
New to interlinear glossing or want a refresher on the notation? These guides cover the standard
and the labels, and link back into the editor.
- Interlinear gloss generator. What the tool does and how to build a gloss in four steps. Start here.
- How to read an interlinear gloss. The three lines, the small-capital labels, and the marks, for beginners.
- Leipzig Glossing Rules explained. The shared standard for morpheme-by-morpheme glosses, with an example for each rule.
- Glossing abbreviations cheat sheet. Case, person, tense, and the notation marks, each with a plain-language note.
- How to gloss your conlang. Leipzig-style glossing for an invented language, with custom scripts and fonts.
- Aligning vertical writing systems. Column layout for Japanese tategaki, Chinese, and Mongolian, and how a horizontal translation sits beside them.
Use it from your AI assistant
There is an agent skill that teaches Claude, ChatGPT, or
a custom agent to build these diagrams from a plain request: ask it to align a translation or
gloss a sentence, and it returns a shareable link. It runs on the same open API, with no key required.
Questions people ask
What is this kind of tool called?
The formal name is a word alignment visualizer. In everyday language people also call it a
word-by-word translation tool, an interlinear-style visualizer, or simply a way to see which
words match in a translation.
Does it handle reordered translations?
Yes. Curved connectors can cross each other freely, so sentences where the target language puts
the verb, subject, or modifier in a different position still look clean. That is one of the main
reasons this visualization is useful for language learning.
Can I align phrases, not just single words?
Yes. Each link joins two word-sized boxes on neighboring lines. You can add several links from
the same word to different partners (one-to-many or many-to-one) by clicking that word again and
choosing another match on the adjacent row. To treat two written words as a single box, for
example a fixed expression, use the join character under Settings → Tokens.
Is this a full machine translator?
No. You type or paste the text yourself; the app does not translate it for you. The value is in
the visualization and the manual control over which words count as matches.
Can I export the alignment as an image?
Yes. PNG, SVG, PDF, and a self-contained HTML file are all supported, along with a shareable
link that encodes every line of text, every connector, and your visual settings.