AI voices for audio drama and radio plays
An audio drama needs a cast: a narrator, heroes, villains and a dozen small parts, all available whenever you rewrite a scene. AI voices make that cast possible for one person at a desk, and every voice sounds the same in episode ten as in episode one.
SuperClonica has narrators, character voices and conversational voices for Supertonic 3, a free online test and a free engine that renders whole scripts on your computer. This guide shows how to cast an audio drama, how to write a script for voicing and how to put an episode together.
Cast your audio drama
Cast voices that differ in pitch and pace, so listeners always know who is speaking without seeing the script.
- Narrators: Conversational Narrator for an everyday tone, Keeper of the Light for legends, Anna for a calm first-person story.
- Leads: Geralt of Rivia, Yennefer of Vengerberg and Triss for fantasy; Sofia, Artyom and Jude for modern stories.
- Villains and dark parts: Wraith King, Visage, Death Prophet, Pennywise.
- Small parts and comic relief: Hobo, Doctor Gnus, Scrooge McDuck, the two heads of Ogre Magi.
- Free voices: Maya, Ciri, Dark Willow, Trevor Philips, Gollum.
Audition voices free online
On the page of any voice, paste a line of 30 to 100 characters, choose the language and press Voice it. In a few seconds you hear whether the voice fits the part and can download the recording. Audition two or three voices on the same line before you cast.
Write a script for voicing
Keep the script as a table: it drives the voicing script, the file names and the order of clips on the timeline.
- One line per speaker per file: CHARACTER, line number, text.
- Stage directions stay out of the text you voice: the engine does not act them out, so describe the emotion in the words or add effects later.
- Ellipses and short sentences for hesitation; a new sentence for a pause.
- Respell invented names and places the way they should sound.
Render and mix an episode
Install the free SuperClonica engine: python -m pip install superclonica-cpu. Render each line with the right voice: superclonica --text "The door was locked from the inside. And yet... it was open." --voice narrator.json --lang en --normalize --output ep01/012-narrator.wav. A small script can read the table and render every line in order. Then place the clips in an audio editor such as Audacity or Reaper, add room tone, music and effects, and export the episode to MP3. The engine runs offline on an ordinary processor, with no character limit.
Your own cast
Character voices are fan-made: use them for fan audio dramas and personal projects within the SuperClonica terms. For an original series you want to publish or sell, build your own cast: train voices in the studio from clean recordings of actors who agreed, or order them on the exchange. Training on your own computer with an NVIDIA graphics card is free.