Make an avatar explain a topic
Goal: have a person from your project's avatars explain a topic, or say words you wrote.
Recipe: Avatar explainer from a topic
Inputs.
- Avatar: a ready avatar of the project with a voice linked to it.
- I will write the words: off, you give a Topic and ReelBolt writes what the person says. On, you type What they say and it is spoken word for word.
- Tone, Language and Shape. Under More options you can change the lip-sync, picture and script models. A workspace admin must have set up a lip-sync model first (see Avatars); without one the run does not start.
Length. The clip is short: about 15 seconds at most. Words that run longer do not fit. This recipe does not make a long presenter video.
What it costs. It uses credits: one still picture of the person plus the seconds of video the speech lasts. The video step has a spend cap. The price shows on the run page before you start. If you write the words yourself, the step that writes them is skipped and not charged.
What you get. One short talking clip of the person.
Permission to send media. The clip step sends the avatar's pictures and voice to the video provider. That is off until you switch on Allow sending your media to AI services in the Permission block above Run on the app page; it applies to the run you start. In the assistant chat, such a run always asks you on its card first. See Avatars.
If it fails.
- The run stops for the voice: link a voice to the avatar in the Avatars tab.
- The words are too long: shorten them to about 15 seconds of speech.
- The clip step does not start: check that sending media is switched on, and that the avatar is ready.