Fish Audio Drama 3
Fish Audio Drama 3: directed voice performances with plain-language directions in square brackets, two-voice dialogue
What it's the best tool for
- Delivery set in plain words in brackets: [nervous, whispering], [excited]
- Two-voice dialogue in a single generation
- 8 ready voices: Laura, Adrian, Sarah, Ethan, Hannah, Jordan, Selene, Egirl
- Speech speed from 0.5x to 2x
- The text language is detected automatically
When to reach for something else
- Up to 3,000 characters per request
- Billing follows text volume: Cyrillic takes about twice as much as Latin
- A dialogue is built from two voices, no more
- Only the 8 ready voices are available
How Fish Audio Drama 3 responds
Fish Audio Drama 3
Four scenarios where it pays for itself
More about Fish Audio Drama 3
Fish Audio Drama 3 Online — Voice with Direction
Fish Audio Drama 3 is a speech model built for directed voice performances. You write direction in plain language in square brackets before a line: [nervous, whispering], [excited], [with a smirk]. A two-voice dialogue comes together in a single request. It runs in the browser on NetRoom.
Direction in plain words
Standard text-to-speech reads a script evenly. Drama 3 performs it: you describe how a line should be said and the voice adjusts. The direction goes in brackets before the line and is never spoken, it only steers the delivery. It is natural language, not a rigid command set: a short [happy] works, and so does a compound [tired, hopeful].
Two-voice dialogue
The model can build a scene with two speakers in a single generation. In the editor, lines are marked with speaker tags and each speaker gets its own voice, so there is no manual track stitching. For dramatic scenes, interviews, and conversational podcasts this saves hours of editing.
8 ready voices
The library has eight voices: Laura, Adrian, Sarah, Ethan, Hannah, Jordan, Selene, and Egirl. The text language is detected automatically, so there is nothing to switch in the settings.
Speed and length
Speech speed is adjustable from 0.5x to 2x. One request voices up to 3,000 characters. For long scripts, split the text into scenes, which also makes direction easier.
Who it is for
Audio-drama and podcast authors get scenes with mood. Game developers and animators get character lines with the right emotion. Marketers get spots where delivery matters. Screenwriters get a quick read-through of dialogue out loud.
How to voice a script
1. Sign up on NetRoom and top up your balance. 2. Open the Sound section and pick Fish Audio Drama 3. 3. Paste the script and put delivery directions in square brackets. 4. Choose a voice, or two voices for a dialogue. 5. Set the speed and download the track. No subscription, billing follows the volume of text, and the current rate is shown on this page.
Another Fish Audio model
If you need 80+ languages and voice cloning, see Fish Audio S2.1 Pro.
Stage your first scene — start on NetRoom.
What changed Fish Audio Drama 3
- + Added voice model Fish Audio Drama 3 (Fish Audio).
Use Fish Audio Drama 3 via the API
The same engine, straight from your code: one key and one balance for text, images, video and sound. Pay only for the requests you make.
curl https://netroom.ai/api/v1/generations \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "fish-audio/drama-3", "input": {"prompt": "A cinematic mountain sunrise"}}'The model id is already in the example. The full parameter reference and prices live in GET /api/v1/models and in the docs.
Try Fish Audio Drama 3
right now
Free access to basic models. No card, no obligations.