To translate a training course into Spanish, first write down which Spanish: es-419 for Latin America or es-ES for Spain. Then export the text, have it translated, have a bilingual reviewer check it against the English and your written spec, rebuild what the export skips (images with text, player buttons, captions, PDFs), replace or subtitle the voiceover, and test a Spanish SCORM package in your LMS. Machine translation can write the first draft; Articulate's FAQ says the accuracy of its own AI translation can vary depending on the language pairs involved.

One English screenshot on slide 14 can move your launch date, because text inside an image is not translated even by Articulate's AI route.

What does translating a course into Spanish include?

The job includes every place a learner meets English: slide text, quiz feedback, player buttons, images, PDFs, narration, captions and the course listing in your LMS. Storyline 360's XLIFF and Word translation process cannot translate closed captions, trigger conditions, player text labels or variable names. Storyline's multi-language route does translate closed captions and the text values used in trigger conditions.

At this time, Articulate's AI translation does not translate text within images or videos, media embedded from outside Articulate 360, or linked or attached files such as PDFs. This page has not read other tools' documentation, so check what yours exports before you set a date. The Storyline and Rise clicks are in how to translate an Articulate Storyline or Rise course.

AssetWhat changesWho does itTypical trap
Slide text and other course textTranslated from the export fileTranslator, then a bilingual reviewerShort labels grow the most and overflow their boxes
Player buttons and menusTranslated outside the exportCourse authorNot in the Storyline export; player text labels are translated by hand in the player
Triggers and variables that test textChecked by handCourse authorTrigger conditions are not exported; translated variable references stop matching
Images, screenshots, diagramsRebuilt in SpanishDesignerNot translated, even by Articulate's AI route
PDFs and job aidsTranslated as their own documentsTranslatorNot translated by Articulate's AI route
VoiceoverRe-recorded, synthetic voice, or kept with Spanish subtitlesVoice talent or text-to-speech, plus a reviewerA recorded track stays in English until you replace it
CaptionsTranslated, or written for the new Spanish audioTranslatorReplacing a video removes its captions
SCORM package and LMS listingA Spanish package, or a language added to one packageLMS administratorSaved progress keeps the language the learner chose

The "Who does it" column is a staffing judgment; the traps come from the sources.

Which Spanish should the course be in?

Use the Spanish your learners use, and name it with a language tag. In W3C's example, es-ES means Spanish as spoken in Spain, whereas es-419 means Spanish as spoken in Latin America. W3C's example does not list countries, so check where your learners' countries fall.

You may be offered "neutral Spanish." No standards body read for this page defines it, so ask what it means.

Spelling is shared. The Asociación de Academias de la Lengua Española groups 23 academies across the Americas, Spain, the Philippines and Equatorial Guinea, and in 1999 all the academies revised the Ortografía jointly for the first time.

In most of Spain, vosotros is the informal plural "you"; in the Americas and in some southern parts of Spain, such as the Canary Islands and western Andalusia, vosotros and its verb forms (calláis, escribisteis) are not used, and ustedes covers the informal plural as well. A course that addresses a crew from Guadalajara, Mexico, as vosotros uses forms the Diccionario panhispánico de dudas says have no use in the Americas.

Write the level of formality into the spec too. For target languages that support grammatical formality, Articulate's AI route offers Formal, Informal or Automatic, and the default, Automatic, picks a level from your content.

W3C advises using a region subtag only if it contributes information needed in a particular context. Your course needs that information, so write es-419 or es-ES.

What do ISO 17100 and ASTM F2575 say about translators?

A translator who knows the subject, then a bilingual reviewer who knows the job, both working from a written spec. ASTM F2575-25 builds the whole job around that spec: specifications agreed in advance and documented in writing, developed in pre-production, followed in production, and used to evaluate the project in post-production.

ASTM F2575 says you, the requester, should run a needs analysis and select a translation grade, and that there can be many correct translations, depending on the specifications.

ISO 17100:2015 sets requirements for the core processes and resources of a quality translation service, and a provider can use it to show conformity. The abstract says raw output from machine translation plus post-editing is outside its scope. So ask what an ISO 17100 claim on your quote covers.

ISO lists the 2015 edition as to be revised, with a new one under development. The public abstract does not list the steps inside, and this page has not read the 19 pages, so ask a provider which clauses they follow.

Machine translation can still write your first draft. Articulate's FAQ says accuracy can vary depending on the language pairs and pairs its AI feature with an in-context human validation workflow; a validator needs only an email address. Your bilingual shift supervisor can be that validator.

What happens to the audio and the images?

A microphone, a speaker and a picture frame beside a course screen, the extras to translate training into spanish

Do you record Spanish narration or add subtitles?

You record Spanish narration, generate it with text-to-speech, or keep the English audio and add Spanish subtitles. If you record or generate Spanish narration, add Spanish captions: W3C's accessibility guidance says captions, which match the language of the audio, are needed for accessibility, and that subtitles in other languages are not directly an accessibility accommodation.

On the subtitle route, W3C notes that subtitles can be a translation of the caption content, including non-speech audio information. Translate the sound cues, not only the dialogue.

In a multi-language Storyline project, text-to-speech narration can be translated automatically, and recorded audio or video can be swapped for a language-specific version. Articulate's guide says to use closed captions if you keep the original media for every language. Replacing a video removes its closed captions, so export them as a.VTT file first and import them afterward.

Do not ship automatic Spanish captions unchecked. W3C says automatically generated captions do not meet user needs unless confirmed fully accurate, and that missing one word such as "not" can make captions contradict the audio. In a lockout procedure, that word is the lesson.

None of the sources read here says when dubbing beats subtitles for training. If your learners must watch hands or a screen while the narrator talks, subtitles ask them to read at the same time.

What do you do with slides, screenshots and images?

Give text room to grow, then rebuild every image with words in it.

The general message is that text will normally expand, but note carefully how the smaller the source message, the higher the likely translation length.

W3C Internationalization, "Text size in translation"

Characters in the English sourceAverage expansion
Up to 10200 to 300 percent
11 to 20180 to 200 percent
21 to 30160 to 180 percent
31 to 50140 to 160 percent
Over 70130 percent

W3C reprints the table as IBM's guidance, and the table is not broken out by language. The copy of the page read for this page is an extract, and its 51 to 70 character row (151 to 170 percent) is higher than the row before it, which cuts against W3C's own point that shorter strings grow the most; it is left out above, so check it on the live page before you rely on it. W3C's advice is to allow text to reflow, avoid small fixed-width containers, and be especially careful about fitting text snugly into graphic designs.

In a multi-language Storyline project, overflowing translated text gets scrollbars, and the Layout Issue Finder lists them.

Then count the images with English in them. Articulate's AI route does not translate that text, so each one is a design task with its own hours.

How do SCORM and your LMS handle a second language?

The plain route is two packages, one English and one Spanish, each its own course in your LMS. Storyline's manual route already works that way: you import the translation into a copy of the project. Which SCORM version to export is covered in SCORM 1.2 vs SCORM 2004.

SCORM 1.2 and SCORM 2004 (4th Edition in the excerpt read) both carry a learner language preference, cmi.student_preference.language and cmi.learner_preference.language, described as the preferred language "for SCOs with multilingual capability". The excerpt of the reference chart read for this page says nothing about packaging two languages in one file.

One multi-language package is possible on Articulate's AI route. Saved progress includes the language the learner chose, so switching means resetting progress, and Rise 360 currently does not let learners reset it themselves in an LMS. Two packages usually mean two completion records in the LMS, so decide before launch which one your compliance report counts.

How do you check the Spanish version before launch?

A bilingual reviewer takes the course as a learner, inside the LMS, with your spec open. ASTM F2575 puts that evaluation in post-production, against the specifications. Fail the quiz once on purpose, so the reviewer sees the feedback for wrong answers too.

CheckWhat goes wrong
Variety and "you" match the specVosotros forms in a course for staff from the Americas
Every slide, layer and feedback screenLonger text overflows or gains scrollbars
Buttons, messages, other player elementsNot in the Storyline export; translated by hand in the player
Quizzes and branches that test textTrigger conditions are not exported
Images, screenshots, PDFsNot translated by the AI route
Captions against the narrationAutomatic captions can miss a word such as "not"
Completion and score in the LMSSaved progress keeps the language the learner chose; switching means a reset

How do you run a Spanish translation project?

You need the authoring files, a translator or a machine draft, a bilingual reviewer who has done the job, someone to rebuild images, and a test area in your LMS. Your authoring tool's export and a spreadsheet for the spec and glossary are the working files; ask your LMS administrator for a test enrollment before you start.

  1. Freeze the English. Storyline flags source text changed after the last translation run, and any change in a text segment retranslates the whole segment; a late edit also means a second review.
  2. Write a one-page spec. es-419 or es-ES, formal or informal address, the plural "you", the audience, and a glossary of terms that stay fixed. ASTM F2575 wants the specifications documented in writing.
  3. List what the export misses. Images with text, PDFs, player buttons, captions, triggers that test text.
  4. Translate. A translator, or a machine draft a translator post-edits; ISO 17100 puts the second outside its scope.
  5. Review against the English. Side by side, marked against the spec.
  6. Rebuild the rest. Images, player labels, captions, audio; export captions as.VTT before you swap a video.
  7. Import and fix the fit, slide by slide and layer by layer.
  8. Publish and test in the LMS with one Spanish-speaking test learner and one failed quiz.
TaskWhoRough time
Spec and asset listCourse owner with a bilingual leadHalf a day
TranslationTranslatorDepends on word count; get the date in writing
ReviewBilingual reviewer1 to 2 times the course's seat time
ImagesDesigner15 to 45 minutes an image
Import, fit and LMS testAuthor1 to 2 days for a 30 minute course

These are working estimates, not figures from any source on this page.

Which Spanish translation mistakes cost the most?

  1. Shipping machine translation nobody reviewed. Articulate says accuracy can vary depending on the language pairs, and W3C's caption guidance shows how one missing "not" can make captions contradict the audio.
  2. Letting expansion break the layout. Short strings grow the most, and buttons are the shortest strings on a slide.
  3. Mixing Spain and the Americas. Vosotros in a course for a crew from the Americas uses forms the dictionary says have no use there.
  4. Leaving English in images and buttons. Player buttons are not in the Storyline export, and text inside images is outside the AI route.

Each one turns up at step 8 and sends you back to step 5, 6 or 7.

Need your course in another language? We build in English, Spanish, French, Arabic, Russian and German. We build the first lesson free from your own material, then give you a fixed price for the rest. Get your free first lesson.

Other questions about Spanish course translation

Should a course be in Latin American or Spain Spanish?

The one your learners use. W3C's example tags are es-419 and es-ES, and one difference that shows up in a course is the plural "you": vosotros in most of Spain, ustedes in the Americas.

Is machine translation good enough for training?

As a first draft a bilingual reviewer checks. Articulate says accuracy can vary depending on the language pairs, and ISO 17100 leaves raw machine output plus post-editing outside its scope.

Do I need Spanish captions if I dub the audio?

Yes, for the Spanish version to be accessible. W3C WAI defines captions as same-language text for the audio.

Can one SCORM package hold English and Spanish?

Articulate's multi-language courses can be hosted in an LMS as one training, and switching language there means resetting progress. SCORM has a language preference for SCOs with multilingual capability, but the reference chart read for this page gives no rule for packaging two languages.