Turn Any Script Into a Talking Video
Write the script. Pick a face. Generate a performance. Go from text to a polished lipsync video in minutes. It's so easy. Get professional quality talking videos that are realistic and convincing in no time at all.
Scripts become performances
Avatar videos, dubbed footage, and ad-ready spokespersons
Hate the Spotlight? You Don't Need to Be in It.
For creators who don't want to be on camera. Use Glif lipsyncing to cut out all the hard part of content creation.
- Talking avatars — Upload a single still photo and generate a realistic speaking performance
- Consistent presenter — Use the same face across every video without ever re-shooting
- Expressive delivery — Add emotion cues to your script to shape tone, pace, and expression
Change the Dialogue Without Re-Shooting.
For global brands and agencies. It's a game changer. Launch the same campaign in ten languages without booking talent ten times.
- Video dubbing — Drop in new audio and sync the lips to match, frame by frame
- Language localisation — Take existing footage and generate synced versions in any language
- Ad spokespersons — Generate a spokesperson from a single photo. Change the script whenever the brief changes.
Create your avatar from scratch and then go from there...
Talking Avatars
Upload a front-facing photo and generate a realistic talking performance. Lips, expressions, and head movement all sync to your audio.
Video Dubbing
Replace the dialogue in any video. Feed in new audio and the lips sync automatically. Clean output, no uncanny valley.
Ad Spokespersons
Generate a spokesperson from a photo. Swap the script for every campaign, market, or audience — no reshoots, no contracts.
Featured workflows
Jump straight in. Pick a card and start creating.
Write the script. Generate engaging performances.
From a single photo to a finished talking video. No camera. No studio. No waiting on talent.
Common Questions
A script or audio file, and either a photo (for an avatar) or an existing video (for dubbing). That's it. The agent handles the sync, the expressions, and the output.
Tips:
Start with clean audio. Slow, clear speech syncs better than anything rushed or mumbled. Read your script out loud first — if you stumble, rewrite it.
Write how people actually talk. Use contractions. Keep sentences under 15 words. The model performs exactly what you give it, robotic script = robotic result.
Use a front-facing photo. Both sides of the mouth visible, even lighting, mouth closed. Side angles and strong shadows will hurt your sync.
Match face to tone. A deadpan face reading an excited script looks off. Pick a photo whose natural expression matches the energy of your content.
Front-facing, well-lit, and close-up. The cleaner the face in the photo, the cleaner the sync. Avoid heavy shadows, extreme angles, or partially obscured faces.
Yes. Upload the source video, provide translated audio or a translated script, and the agent generates a synced version with lips matched to the new dialogue.
Add emotion cues directly in your script — words like 'said firmly' or 'whispered' or 'with excitement' shape how the performance is delivered. Iterate on the first few seconds first before generating the full video.
Yes. You own full rights to everything you generate. Use it across Meta, TikTok, YouTube, LinkedIn, or any other platform without licensing restrictions.
Yes. Upload the same photo each time and generate as many scripts as you need. The result is a consistent on-screen presence across every video — without ever filming again.
