The animation engine for agents: blazingly fast math and 3D videos with Manim's Python API, on a Rust GPU core.
title: "Voice" description: "A voice that explains: say a line while animations play during its words, choose a voice, read captions, or make a voice of your own."
Voice
import manimgx as m
class VoiceHero(m.Scene):
def construct(self) -> None:
square = m.MathTex("x^2", font_size=96).shift(2 * m.LEFT)
derivative = m.MathTex("2x", font_size=96).shift(2 * m.RIGHT)
arrow = m.Arrow(square.get_right(), derivative.get_left(), buff=0.4)
script = m.Text(
"The derivative of [x squared] is [two x].", font_size=32
).to_edge(m.DOWN)
self.add(square, script)
self.play(m.Indicate(square))
self.play(m.GrowArrow(arrow), m.Write(derivative))
self.wait()
[say][manimgx.Scene.say] speaks a line with the scene's voice, and plays animations while it speaks. Put the words of each animation in brackets, and give the animations in the same order: each starts with the first of its words. The scene waits until the line ends, so lines don't overlap; change the words, and the animations follow the new words.
The scene's voice speaks through a text-to-speech service. ManimGX keeps what it says in a
folder named voice/, next to the scene's file, so the scene renders again without the
service, and only a changed line goes to it again. What the voice says becomes the video's
captions.
::: manimgx.Scene.say options: heading_level: 2
::: manimgx.Scene.speech options: heading_level: 2
::: manimgx.Scene.voice options: heading_level: 2
::: manimgx.Speech options: heading_level: 2
Voices
The voices ManimGX brings are fal.ai's: m.Fal, with the model you choose and its
settings. Your editor shows each model's settings.
::: manimgx.Fal options: heading_level: 3
Models { #manimgx.audio.fal }
::: manimgx.audio.fal.ElevenV3 options: heading_level: 4
::: manimgx.audio.fal.MiniMax options: heading_level: 4
::: manimgx.audio.fal.MiniMaxVoice options: heading_level: 4
::: manimgx.audio.fal.MiniMaxAudio options: heading_level: 4
::: manimgx.audio.fal.MiniMaxLoudness options: heading_level: 4
::: manimgx.audio.fal.MiniMaxTimbre options: heading_level: 4
::: manimgx.audio.fal.MiniMaxPronunciation options: heading_level: 4
::: manimgx.audio.fal.Gemini options: heading_level: 4
::: manimgx.audio.fal.Inworld options: heading_level: 4
::: manimgx.audio.fal.Qwen options: heading_level: 4
Captions
::: manimgx.Scene.add_subcaption options: heading_level: 3
A voice of your own
A voice is any function from a text to a [Speech][manimgx.Speech]: a sound that knows its words, and when each is said. Any text-to-speech service can be one.
::: manimgx.Voice options: heading_level: 3
::: manimgx.Word options: heading_level: 3
::: manimgx.timed options: heading_level: 3
::: manimgx.estimate options: heading_level: 3
::: manimgx.cached options: heading_level: 3