How to make an AI cartoon with different voices for each character

Choose character dialogue or a narrated cartoon

A narrated cartoon has one voice telling the story over its pictures. A dialogue cartoon is carried by the cast: each scene has a named speaker and that character’s voice. Choose Characters talk when your script depends on an argument, conversation or exchange of jokes.

A cast with several pictures does not automatically make a script multi-voice. Assign the speaker for every line. For an initial attempt, two characters are enough to check the complete workflow without managing a large cast.

Build a cast that looks and sounds distinct

Open AI Story and switch to Characters talk, or start from Fruit videos for a ready cast. Add characters from the library or prepare your own pictures in the selected look. Keep a stable name and a short appearance description for each character.

Choose a voice for each speaker and audition the same sample sentence in both. Listen for useful differences in pace, tone and clarity rather than choosing two voices that only look different in the picker. Keep the assignments fixed as you generate the scene sequence.

Write one speaker’s line per scene

Paste the example dialogue after adding Milo and Pip to the cast and assigning their voices. Parentheses after a name describe the scene; the words after the colon are spoken. Keep the button visible when the joke depends on it.

Characters talk supports up to 40 scenes, with a maximum of 240 characters per spoken line. Shorter lines are usually easier to stage and revise. The example below is a script to try, not a transcript of the separate finished video sample.

Check voices before making the characters talk

Generate the pictures and voice lines, then listen to the conversation in order. Check that a line belongs to the correct speaker, names are pronounced clearly and the delivery makes the exchange understandable. Shorten a rushed sentence or change the voice assignment before producing the talking clip.

Use delivery cues sparingly and audition the result. A whisper may suit one reveal; adding dramatic tags to every line can obscure the meaning. These cues influence performance but still need a listening pass.

• Every scene points to a character in the cast.
• Every speaking character has a selected voice.
• The visible speaker matches the line.
• Short pauses support the joke or emotional turn.
• The voices remain distinct at a comfortable volume.

Make talking clips, then finish the conversation

Use Make them talk for the approved scenes and review the displayed AI credit cost first. Each clip has one active speaker; the format is an alternating conversation, not two characters speaking over one another.

Watch the finished sequence with sound, then inspect the captions at phone size. Fix incorrect names and captions that hide the mouth or important prop. Keep music behind the dialogue and recheck transitions after any timing changes.