How do you make two animated characters talk to each other?
Three moves, all in the browser:
- Add a second character. In the Character panel, add an extra character and give it its own look — outfit, colors, props — so the two are instantly tellable apart. Position them facing each other on the canvas.
- Open Dialogue mode. It lives in the Scene panel, and it swaps the single prompt box for a conversation: a list of lines, each one assigned to a character.
- Write the exchange and generate. Each line is voiced for its character and the speakers animate in turn — the whole conversation plays as one continuous scene.
Does each character get their own voice?
Yes — every line is spoken as the character it belongs to, so the exchange sounds like two people rather than one narrator doing both parts. The same performance tags that work in solo voice-over work per line here: a [pause] before a reply is the difference between an answer and a guilty answer, and wrapping one retort in <loud> turns a chat into an argument.
How does the camera know who is talking?
Dialogue mode has a focus-on-speaker option: the camera moves to whoever owns the current line, back and forth as the conversation flows. It's the single setting that makes a two-character scene feel directed instead of surveilled — a static wide shot of two figures reads like security footage; a camera that leans toward the speaker reads like a show. Turn it off when you want a fixed frame, for example when the joke depends on seeing both reactions at once.
Can characters move and gesture while they talk?
Yes, and this is where conversations come alive. Each line can carry its own motion prompt alongside the spoken text: the excuse comes with a shrug, the accusation with a step forward, the defeat with a slow slump. If you leave a line's motion empty, the character still performs the line naturally — but the strongest scenes choreograph the big lines deliberately. Body language is half the dialogue.
How do subtitles work in a conversation?
Each line's subtitle appears while that line is spoken, automatically timed to the audio. Combined with the two characters' distinct looks, a viewer on mute can still follow exactly who said what — which matters, since conversational clips are precisely the kind of content people watch in feeds with the sound off.
Can I fix one line without regenerating the whole conversation?
Yes. Dialogue has the same half-regeneration controls as the main generator: rewrite the weak line, then re-run only the voices (keeping all the motion) or only the motion (keeping all the recorded voices). Everything that already worked is reused, so polishing a conversation is iterative rather than a gamble on a full re-roll.
What kinds of scenes work best?
- The skit: setup line, escalating middle, one-line punchline — the classic two-hander, and the easiest to write.
- The interview: one character asks, the other answers; swap the answers each episode and you have a series.
- The explainer duo: a "smart one" explains, a "confused one" asks the questions your audience would — the oldest teaching trick in animation, because it works.
Read your dialogue out loud once before generating. Lines that feel natural spoken generate better rhythm than lines written to be read — and cutting one exchange almost always improves the scene.