01
Define the localized deliverable
Write down whether you need a translated script, subtitles, a dubbed audio track or a finished video with synchronized speech. Record speaker count and target language pairs.
Why this matters: Different services package transcription, translation, timing, video rendering and audio export differently.
02
Create and review the source transcript
Check names, numbers, timing and speaker boundaries in the original transcription. Correct source errors before translating.
Why this matters: Automated translation cannot reliably fix a mistranscribed sentence.
03
Review translation and voice permissions
Have a fluent reviewer check tone, terminology and cultural fit. Confirm consent to reuse any identifiable voice, face or likeness in synthetic output.
Why this matters: A feature being technically available does not grant you the rights to clone a voice or republish someone else's footage.
04
Inspect timing and deliver
Render a short test segment, check lip/timing drift, speaker changes, captions and exported quality. Then finalize the full project and verify any watermark or plan limits.
Why this matters: Language length and pacing can differ; errors may only appear once audio and video are combined.