After Effects text to speech: an AI voice that drives your animation
After Effects animates to audio very well, but it does not turn text into speech. For an explainer, a kinetic typography piece or a talking character, the practical way is to voice the script with a text to speech tool, import the audio and let After Effects read its loudness.
The free SuperClonica engine voices the script on the same computer, offline, with natural voices and character voices from the catalog. This guide covers the whole path from script to a composition that moves with the voice.
Voice the script
Install the engine once with python -m pip install superclonica-cpu; the model downloads on the first run, and after that everything works offline. Save the script as a text file and voice it: superclonica --text-file script.txt --voice F3 --lang en --normalize --output voice.wav. For a piece with several scenes or characters, voice each line into its own file, so a change means redoing one file. To choose a voice, try a line on its catalog page first, free and with no account.
Import and place it
In After Effects choose File, Import, File, or drag the WAV into the Project panel, then drop it into your composition. Press L twice on the layer to see the waveform and line up text and shots with the words. Preview with audio from the Preview panel, and lower any music layer under the voice with its Audio Levels.
Make things move with the voice
Select the voice layer and choose Animation, Keyframe Assistant, Convert Audio to Keyframes. After Effects adds an Audio Amplitude layer whose Both Channels slider follows the loudness of the voice. Link any property to it with an expression, for example the scale of a mouth or a speaker icon: thisComp.layer("Audio Amplitude").effect("Both Channels")("Slider"). For real mouth shapes that match the sounds, our lip sync guide shows how to turn the same WAV into timed visemes.
Fix and polish
If a line sounds wrong, change the text and re-voice that file instead of editing the waveform: it takes seconds, and the composition picks up the new file when you replace the footage. Commas and full stops add pauses, and numbers are read most reliably as words. Every voice speaks 31 languages, so the same voice can narrate versions of the animation in Spanish, German or Japanese; Chinese is not supported yet.