HeyGen's Video Translation tool makes it easy to localize your content. You can re-voice any video into a new language while keeping the original speaker's exact voice and perfectly syncing their lip movements. Simply upload your video, choose your target languages and translation engine, and hit Translate.
Getting Started
Before configuring any language settings, you first need to import your video.
You have three options to add a video:
Browse local files: Drag and drop your video into the designated area, or click to upload a file from your device (MP4, MOV, or WEBM format, up to 5 GB).
Select an existing project: Choose a video you've already created within your HeyGen workspace.
Paste a URL: Paste a direct link from YouTube or Google Drive.
Once your video is imported, you can configure your translation settings. Select the following:
Source Language – the language spoken in the original video. Use Auto Detect or choose manually. Your source video must be a single language — videos with multiple spoken languages will produce a failed translation or an inaccurate transcript.
Target Language – the language to translate into. You can select up to 10 languages per job.
Base language vs. Localized variant: Many languages offer both options.
Base (e.g. "English") — preserves the original speaker's voice identity and accent. Best for interviews, documentaries, reaction videos, and livestreams where the speaker's character matters.
Localized variant (e.g. "English – American")- also known as accents or dialects- this prioritizes a native target-language accent. Best for tutorials, onboarding, demos, and corporate content.
🛎️ Pro tip: The default accent for each language is the most common regional variant — for example, American English is the default for English.
The input speaker's voice is always used for voice similarity regardless of which you select.
We recommend translating into one language first to confirm quality before running all languages at once — this avoids spending credits on translations that may need revision.
After selecting your languages, choose a translation engine:
Audio Only = 2 credits/minute
Translates and re-voices the audio only
No lip sync
Fastest and most cost-effective option
Best when lip synchronization isn't required
Speed = 5 credits/minute
Includes natural lip sync
Optimized for front-facing videos
Best for:
Speakers facing the camera
Minimal facial occlusion
Single speakers or simple conversations
Lower credit cost with excellent quality for standard videos
Precision = 10 credits/minute
Includes the highest-quality lip sync
Better handles:
Speaker switches
Camera angle changes
Side-profile shots
Facial occlusions (hands, microphones, etc.)
More complex conversations
Choose Precision when quality and lip-sync accuracy are the priority.
Precision mode costs double the credits of Speed mode. This reflects the advanced AI processing required to deliver lip-syncing even in the most challenging video conditions, like side-profiles and heavy movement.
Use Speed if: the video is mostly front-facing (within ~45°), has little to no facial occlusion, single speaker or simple conversations, and doesn't require the highest lip-sync fidelity.
Use Precision if: the video has significant side-profile views, facial occlusions (hands, microphones, fade-in/fade-out), complex multi-person conversations, or you need the highest translation and lip-sync quality.
💡 Note: Lip-sync is controlled by your engine choice rather than a separate on/off toggle. Selecting Audio Only turns lip-sync OFF, while selecting Speed or Precision turns lip-sync ON. There is also no translation prompt, tone, or context field.
Other tools to improve translation quality:
Brand Glossary — keeps brand names and product terms consistent across translations.
Proofread (also called "Review & Edit") — review and edit the translated script word-by-word before generating the final video. Available on Pro, Business, and Enterprise plans.
Add Transcription (SRT) — upload an SRT file to guide the translation. Important distinction:
Input SRT (uploaded before translation): HeyGen uses it as reference but re-translates the content — the output will not match your file word-for-word.
Output SRT (uploaded in Proofread mode): spoken exactly as written. This only works in Proofread mode — it cannot be used for direct translation.
Advanced Settings
Click Advanced on the translation setup page to access these options:
Brand Glossary: Maintain consistent pronunciation and translation of key terms (brand names, product names). (Learn more)
Enable Dynamic Duration: Allows segment durations to stretch or compress by up to ±20% to create more natural-sounding translated speech. Enabled by default. Note: This may affect the total length of your video.
Create a Collection: Groups all multilingual translations into one organized collection on your Projects page.
Remove Background Sound: Removes background music from your video. This is an optional, user-controlled setting. Note: Sound effects (e.g. dog barks, car horns) are always removed from the translated audio regardless of this setting.
Output to Original Video Specs: Retains your video's original resolution (720p, 1080p, 4K). Available on Pro, Business, and Enterprise plans only.
Enable Captions: Adds translated subtitles to your video. When enabled:
Your output includes two versions — one with captions burned into the video, and one without.
Caption files export as .srt only (no VTT, ASS, or custom styling options).
Enhance Voice: Boosts volume and improves clarity for a studio-quality sound. Note: This may slightly alter the original tone of the audio.
🛎️ Enterprise note: Enterprise users can upload local voice files to create a more accurate voice clone for a specific video.
Once you've set everything, click Translate. You'll receive an email notification when your translation is complete — you don't need to keep the page open.
What Gets Translated (and What Doesn't)
Translated: The spoken audio — re-voiced in the new language in the original speaker's voice — and, optionally, lip movement and on-screen captions.
Not translated: Text baked into the video image (on-screen graphics, hard-coded captions in the source). Background music is left unchanged unless you enable Remove Background Sound.
Singing and music: Sung lyrics are translated for the most part.
Unsupported source languages fail instantly and use no credits.
Video Translation File Requirements
When using HeyGen's Video Translation tool, your source video must meet specific requirements. These details can also be found under the "Requirements" info button within the tool.
Video file requirements:
File type: mp4, mov, webm
File size: up to 5GB
File length: 2 seconds to 10 hours (depending on your plan's max duration limit)
Google Drive Link: file must be shared with "anyone with the link" access
YouTube Link: use the full URL, not a shortened one
Vimeo Link: use the full URL, not a shortened one
Translation Best Practices
Follow these guidelines to get the best translation output.
Camera Angle
Ensure speakers face the camera directly or at no more than a 45-degree angle.
Proximity to Camera
Keep speakers within 10 feet of the camera for clearer audio and better facial capture.
Single Speaker
Only one person should speak at a time to improve lip sync accuracy.
Shot Composition
Use close-up shots and minimize camera cuts for smoother results.
Background Noise
Reduce background noise as much as possible to improve audio quality.
Sharing + Downloading Your Translated Video
Once your translation is complete, click into the project to share or download.
Download: If you enabled captions, you'll have two download options — with captions burned in, and without.
Share link: Generates a direct link viewable by anyone with the URL. You can add password protection via the share settings (this is the same share feature used across all HeyGen videos).
Audio-only download: Available on Enterprise plans only.
SCORM / xAPI export: Export as a SCORM package for use in an LMS. Available on Business and Enterprise plans. See the SCORM Export guide.
Translation Pricing
All paid plans include a credits balance that can be used for video translation.
Your Credits balance = your Video Translation capacity
Credit cost by engine:
Audio Only: 2 credits/minute
Speed: 5 credits/minute
Precision: 10 credits/minute
Examples by plan (Speed engine):
Creator: 600 credits → 120 min VT
Pro: 1000 credits → 200 min VT
Business: 1500 credits → 300 min VT
If you need more translation minutes than your plan allows, you can upgrade your plan or purchase additional credits.
Frequently Asked Questions (FAQ)
What are the video specs / requirements for downloading?
Free users can export up to 720p, Creator up to 1080p, and Pro/Business/Enterprise can export in 4K if the original video is the same resolution.
My video exceeds my plan's duration limit — what can I do?
Use the Trim Video option before translating to shorten your video to fit within your plan's limit. Alternatively, split the video into parts and translate each separately.
I have enough credits to translate more than 30 minutes, why did I hit a 30-minute limit?
Creator and Pro plans cap the max duration per video at 30 minutes. Credits add translation minutes to your pool but don't change the per-video cap. You can translate longer content by splitting the video into parts, or by upgrading to Business (60-minute limit) or Enterprise (no limit).
Why did my translation fail instantly?
If the source language is not supported, the translation fails immediately and no credits are charged. Check that your source language is on the supported input languages list above.
Why did my translation fail?
Translation can fail due to improper speaker positioning, too many visible speakers, frequent camera cuts, or background noise. See the Translation Best Practices section above.
My translation failed due to moderation. What can I do?
Submit for manual review by emailing [email protected] or using the chat feature. Include the video ID so our team can review it quickly.
Why are all my translations going into a folder when I generate multiple at once?
When translating into multiple languages in one job, HeyGen automatically creates a subfolder to keep your Projects view organized. Single-language translations appear directly in your main Projects folder.
My translation is inaccurate — what can I do?
AI translation can't guarantee 100% accuracy. Please reach out to our Support team as well so we can help and recharge some of your credits!
For higher accuracy and control for each word in your script, use Script Proofread (Review & Edit) — available on Pro, Business, and Enterprise plans — to adjust the translation before finalizing.
Why is the lip-sync off or not accurate?
See the Translation Best Practices section above. Consider switching to Precision mode for more complex videos.
Text on my screen didn't translate — why?
Video Translation only translates spoken audio (and optionally lip movement). On-screen text, graphics, and hard-coded captions are not translated. For on-screen text translation, create a video in HeyGen Studio instead.
Can I use the API to translate?
Yes. The HeyGen API supports video translation for developers who want to automate or integrate translation into their own workflows. See our HeyGen API Plans article for more details.