Mixer AI › Knowledge Base › How to › How to create a talking portrait

How to create a talking portrait

Prepare two things: a clearly visible face and finished speech audio. The avatar model animates the image to the recording, so approve the script and voice before starting.

Try it on Mixer AI

Step by step

Choose a frontal or near-frontal portrait. Keep hair, hands and props away from the mouth, with some space around the head.

Record speech or synthesize it in ElevenLabs. Remove unnecessary noise and check names, pronunciation and pauses.

Upload the photo and audio to Kling AI Avatar. Test a short passage first, checking lips, chin, head turns and the line ending.

Which model to use

Kling AI Avatar

A speaking or singing portrait from one photo and a finished audio track.

ElevenLabs

Voice a script with one speaker or build a dialogue with a separate voice for each line.

FAQ

Can I enter text instead of audio?

Kling AI Avatar requires a finished audio file. Synthesize text in a separate tool first.

Can I create a singing character?

Try a track with clear vocals. Multiple voices and loud accompaniment make synchronization harder.

How do I make a two-person conversation?

Prepare and process each speaker separately, then assemble the lines in an edit.

Try it on Mixer AI