HeyGen Avatar V: a presenter video from a voice recording

You have a voice-over and no one to stand in front of the camera. HeyGen Avatar V fills that gap: you choose one of its ready-made presenters, upload a speech recording, and it returns a video of that presenter saying your words, lip-synced, at up to 4K. There is no prompt box and no face to upload. Two inputs, a few settings, and the whole job is in getting those right.

Step 1: choose the presenter by name, pose and angle

The avatar list holds hundreds of entries, and their names read like a shot list: a first name, then whether the person sits or stands, then the setting (an office, a sofa, a kitchen, a yoga mat, a picnic), then the camera angle, front or side. So 'Annie, office, sitting, front' and 'Annie, office, sitting, side' are the same presenter in the same room from two cameras.

That makes a cheap two-camera edit possible. Render the same recording twice, once on the front version and once on the side version of the same presenter, and cut between them on the sentence breaks. It looks like a filmed interview instead of one locked shot.

Step 2: prepare the audio

The audio drives everything: the mouth, the pauses, the rhythm of the head. A clean recording in a quiet room beats a perfect script read in an echoing kitchen. If you would rather not record your own voice, make one with a text to speech model from the models page and feed the file in. Write the script for the ear: short sentences, a pause where a slide would change, numbers spelled the way you want them said.

Step 3: pick the frame for where it will play

Aspect ratio comes in 16:9 (the default), 9:16, 1:1, 5:4 and 4:5. Decide by the destination, not by taste: 16:9 for a website or a presentation, 9:16 for short vertical clips, 4:5 or 1:1 for a feed. The aspect ratio guide has the longer version of that choice.

Resolution is 720p, 1080p (the default) or 4K. 1080p is right for nearly everything people watch on a phone or in a browser. Save 4K for a big screen or for footage you will crop into later.

Step 4 (optional): swap the background

Under the advanced settings you can add a background image. Each avatar already comes with its own room, so you only need it when the room is wrong for your brand: your shop, your office, a plain colour behind a tutorial. Use a picture with the same camera height and a little blur, or the presenter will look pasted on.

Avatar V or one of its neighbours?

• Avatar V: a stock presenter, your audio, up to 4K. Best when you need a person on screen and do not mind that it is not you.
• Avatar IV: your own portrait photo instead of a stock presenter, with written directions for expressions. The Avatar IV guide covers it.
• Avatar V Digital Twin: works from text or audio, and can remove the background for a transparent video.

What it will not do

You cannot bring your own face here, and you cannot direct the performance: there is no prompt, so the gestures and expressions come from the audio and the chosen avatar alone. If the presenter must look surprised on one particular word, record the word with that energy, or pick a model with a directions field. For other ways to get a person moving on screen, see choosing an AI video model.