When choosing an e-learning video captioning provider for multilingual training, require accessible source-language captions, reviewed subtitles in each target language and files that work in your learning platform. Before a library-wide rollout, give candidates the same clip and deliverables list. Load their files into your LMS: a transcript alone will not tell you how the lesson works for employees.
Define the video captioning job before requesting bids
Send candidates a list of videos with durations, source and target languages, delivery dates and the LMS or player employees will use. Flag multiple speakers and noisy audio. Call out screen recordings, technical instructions and text embedded in the picture, too; a basic transcription quote may not cover the extra work.
Define the output as well as the languages. Source-language closed captions should convey meaningful speech and audio information to viewers who cannot hear the soundtrack. Translated subtitles carry the spoken message into another language; if learners also need speaker changes and meaningful sounds in that language, include those cues in the order. A written transcript helps with reference and review, while timed captions let learners follow the video as it plays.
|
Request |
Why it matters to the training team |
What to ask for |
|---|---|---|
|
Source-language captions |
Access to speech and meaningful audio without sound |
Speaker identification where needed, sound cues, punctuation, synchronized timing |
|
Translated subtitle tracks |
Comprehension for employees reading another language |
Named target languages and locales, review by a qualified linguist, approved terminology |
|
Transcript |
Search, reference and alternate access to spoken material |
Editable text with agreed speaker labels |
|
Visual-text handling |
Screen labels or warnings may remain untranslated |
A separate list of on-screen text requiring translation or video editing |
Quizzes, PDF handouts, the interface and assessments need their own translation scope. Price them separately if they are part of the project, so you can compare the captioning quotes on equal terms.
Evaluate accessibility and translated subtitles separately
Test captions against what the learner actually hears
Play a scene with a speaker change, an important sound and an instruction over a screen demonstration. A viewer reading the captions should know who speaks and what meaningful sounds occur, without text obscuring the action. Listen for missed negations, names and numbers; then look at punctuation and timing. Automated speech recognition can provide a first pass, but someone must catch words that change an instruction.
For prerecorded video, WCAG 2.2 success criterion 1.2.2 calls for captions for prerecorded audio in synchronized media, with stated exceptions. Your supplier can prepare a track; your team must still assess the published player and the rest of the training experience against your accessibility requirements. Inspect the finished lesson rather than relying on a “WCAG compliant captions” claim.
“Captions are provided for all prerecorded audio content in synchronized media, except when the media is a media alternative for text and is clearly labeled as such.” – W3C, WCAG 2.2
Ask who reviews each translated track
Have the provider translate from an approved source transcript, revise against the video and check the final timed track. A product name, safety term or acronym can sound plausible yet be wrong in the lesson. Share a glossary with definitions, preferred translations and terms that must stay in the source language. Find out how the team flags ambiguous speech.
Approve the source transcript before translation starts. If you change a policy statement or safety instruction after the vendor has timed several languages, have them identify the affected cues and return updated tracks for review.
Have a reviewer watch the complete sample in the target language. The lines must leave time to follow a demonstration, break without changing meaning and match the on-screen action. Watch for confusing speaker changes. Even a correct translation becomes hard to use when a single cue holds a paragraph.
Decide where AI fits the workflow
AI may draft the transcript or translation before a person corrects errors and approves the video. Sensitive, specialized material may call for a fully human workflow. In either case, find out who checks terminology and timing, whether a target-language reviewer sees the video, and who handles corrections after sign-off.
Require LMS-ready files, not just a preview link

Your player determines the handoff format. Common sidecar options include WebVTT (.vtt) and SubRip (.srt); the W3C WebVTT specification describes timed text tracks associated with audio or video. Request a small export in the format your platform accepts, then upload it to a staging course before ordering the full library.
Request one named track per language, editable files and a manifest tying each track to its video version. Include locale identifiers if your LMS distinguishes language variants. Test accented characters and a non-Latin script if your project uses them. On the devices employees use, try turning captions on and off and selecting each language. The text should stay readable over the video.
|
Handoff issue |
Evidence to request in the sample |
|---|---|
|
File and encoding |
An actual |
|
Language selection |
Separate, clearly labeled tracks visible in the learner-facing player |
|
Version control |
Video filename or ID, language/locale, revision and matching track name |
|
Corrections |
Editable timed-text file and a documented path for replacing an outdated track |
|
Burned-in text |
Clarification on whether the quote covers only captions or also changes to the video image |
If you publish through an authoring tool, test its exported package too. Tracks may work in a standalone player yet fail after packaging or when the host platform attaches them differently. Use the learner-facing lesson for acceptance, rather than the vendor’s proofing page.
Run one sample that exposes real training risks
Send each shortlisted provider the same clip and brief. Include a speaker transition, a technical term, a number or warning, and a screen label the narration refers to. Supply the approved glossary, target language and locale, an example of the LMS player, and required file formats. Remove confidential information or use an approved secure transfer route.
Request a source-language caption file, one translated subtitle file, unresolved-term notes and the review steps applied to the sample. Upload the files and watch them in context instead of grading a styled demo.
Use the same scorecard for every candidate:
|
Criterion |
Pass condition |
|---|---|
|
Instruction accuracy |
Negations, warnings, units, names and numbers match the video |
|
Caption access |
Speaker changes and meaningful sounds are conveyed when needed |
|
Translation |
Approved terms stay consistent and the instruction retains its meaning |
|
Viewing experience |
Timing, line breaks and placement leave the demonstration legible |
|
LMS acceptance |
Each track imports, displays correctly and can be selected by a learner |
|
Issue handling |
The vendor identifies unclear speech and fixes documented errors |
Log errors with timestamps, severity and proposed fixes. A phrasing preference carries less risk than a wrong safety step or a track that will not load. Buy enough review for your highest-risk lesson types.
Request an itemized quote you can compare
Ask each candidate to price the same scope by video minute, source-language captioning, target language, review level, formatting, platform testing and revisions. Put minimum charges and rush fees on separate lines. Make sure the quote says whether it includes transcription, translation from an approved script, timing, sound cues, target-language review and your required formats. Otherwise a low per-minute rate can leave out work you need.
Plan for updated footage when a policy, product or interface changes. Compare the cost of replacing a sentence with the cost of re-timing or retranslating a lesson. Agree on review rounds, error corrections, terminology approval and ownership of the editable files.
For confidential employee or product material, cover file transfer, access, subcontractors, AI processing and deletion after delivery in the service agreement.
Copyable request for proposal: “Please quote source-language accessible captions and reviewed translated subtitles for , in [languages and locales]. Deliver separate editable [required formats] tracks named to match each video, include terminology review against our glossary and a playback test in [LMS/player]. Show turnaround, review stages, revision allowance, minimums and any extra charges for on-screen text or video changes. Please return a sample using our supplied clip.”
Make the choice on evidence
Choose the provider whose sample preserves instructions, gives learners usable captions and loads in your LMS. The quote should name who reviews and revises the files. If the timed-text files fail or the scope is unclear, request a corrected sample before scaling up.
Team Stream offers captioning, subtitling and translation for teams delivering multilingual, accessible communication. Share a training clip, your terminology list and LMS requirements with Team Stream, and request the same sample and itemized quote you ask of other providers. You can then choose the workflow that works for your learners and your publishing process.