What does Drama TTS mean?
Drama TTS is text-to-speech that performs dialogue from written directions, so a line can be whispered, sneered or shouted without recording an actor. Fish Audio Drama 3 is a model built for it.
Fish Audio released Drama 3 on 13 October 2026 as a text-to-speech model built for directed performances. In DramaSo it is a selectable voice engine: write [sneering, low voice] before a line and Drama 3 plays it that way. Every line of your episode is voiced, then stitched into one track.
40
welcome credits
1080p
export
9:16
native vertical
Try it here — no sign-in
Three sample episodes, broken down the way DramaSo builds them: timed beats, a voiced cast and storyboard frames. Tap a card, then make your own version in the studio.
Mocked for three years as the family's live-in son-in-law, he lays one black card on the banquet table — and the room goes silent.
Beat breakdown
Cast & voices
Storyboard frames
Sample cut
Sample breakdowns from DramaSo's showcase — nothing is generated or charged on this page.
She signs a 99-day contract marriage with a cold CEO to save her father's company. He was never supposed to fall first.
Beat breakdown
Cast & voices
Storyboard frames
Sample cut
Sample breakdowns from DramaSo's showcase — nothing is generated or charged on this page.
Betrayed by her fiancé and her sister, she dies — and wakes up on the morning of the same wedding.
Beat breakdown
Cast & voices
Storyboard frames
Sample cut
Sample breakdowns from DramaSo's showcase — nothing is generated or charged on this page.
From script to film
Every stage below is a real DramaSo output. Words are cheap, so you review the hook before any video renders.
Paste a script or one line. Parsing into characters and shots is free.
Hook, conflict, turn, climax and cliffhanger — timed to the second.
Every character gets a distinct AI voice; hear the whole episode first.
Each beat becomes a 9:16 frame you can swap before spending on video.
Dialogue, effects and music mixed into one episode — add picture any time.
Free to start with welcome credits. Plans and credit pricing are on the DramaSo homepage. See pricing →
Drama 3 makes a line sound acted, but a short drama also needs a script, a consistent cast, shots and a vertical edit. DramaSo writes the beats, keeps every character's voice description, dubs the episode with the engine you pick and turns the same project into a 9:16 video.
Frames and cuts from DramaSo's showcase — the look your episode starts from.
Most TTS engines either ignore stage directions or read them out loud. Drama 3 treats the bracket as a note to the actor.
In our 14 October 2026 run, `[sneer, low voice] You thought hiding here would save you?` came back as the line only, delivered slowly and under the breath. The transcript did not contain the cue.
If a character has no Fish voice yet, DramaSo puts the character's voice description in the same bracket, so a ‘deep, commanding young man’ sounds like one without picking a preset.
When a Drama 3 request fails, DramaSo removes the bracket and voices the line with Fish S2.1 Pro, so the episode track has no gap.
Drama 3 sits in the same voice list as Doubao Seed, Gemini 3.8, Eleven v4, Eleven v4 Turbo, Qwen TTS, Fish 2.1, HeyGen Voice and HeyGen Voice Turbo. Switch the engine, re-dub, listen again.
Eleven v4 keeps the direction in parentheses. Fish S2.1, Gemini, Qwen and HeyGen Voice read cues aloud, so DramaSo sends them the dialogue only.
Gender and voice description stay on the character, and each engine maps them onto its own voice slot. Changing engines does not mean re-casting.
Episode dubbing costs 1 credit for every five lines, so you can hear the whole episode in two or three engines before you render any video.
Nothing to install. You only sign in when you generate.
Paste a finished script or a one-line premise. DramaSo splits it into lines, each with a speaker and a direction.
Open the voice step and choose Fish Drama 3. The character's voice description and the line's direction go into the bracket.
Listen to the stitched track. Re-dub with Eleven v4, Gemini 3.8 or HeyGen Voice to hear the same cast on another engine.
Creators whose dialogue has to carry emotion without a recording session.
Sneers, whispers and breaking voices on the hook line, written straight into the script.
Ship the episode as audio drama first, then add pictures to the same project later.
Dub one episode with several engines and keep the one your audience reacts to.
Tools that pair with Drama TTS.
Drama TTS is text-to-speech that performs dialogue from written directions, so a line can be whispered, sneered or shouted without recording an actor. Fish Audio Drama 3 is a model built for it.
Put the direction in square brackets before the line, for example `[nervous, almost whispering] I saw him.` Drama 3 uses the bracket for the delivery and does not speak it.
FAQ'S
Ask us anything!