Voices

In the voice picker you can filter by type: Dramatic, Natural, Voice changer, and Cheap (coming soon), plus Record It Myself at the top of the list. Each gives you a different level of control over how your scene partner sounds during rehearsal, and every voice shows its per-minute rate so you know what it costs before you use it.

Can you direct it?

The biggest difference between voice types is whether you can direct the read. Some voices take direction; others give a fixed, clean read. Here is the quick map:

Voice type	Who performs it	Can you direct it?	How you direct it
Dramatic	AI	Yes	Text speech instructions per line
Natural	AI	No	Fixed clean read, nothing to set
Voice changer	You, with an AI voice on top	Yes	Your own performance (record or upload)
Record It Myself	You	Yes	Your own recording or uploaded file
Cheap (coming soon)	AI	Coming soon	Built for low-cost line learning

Rule of thumb: to shape the emotion, use Dramatic (direct it with words) or Voice changer / Record It Myself (direct it with your own performance). For a clean, consistent read for a self tape, use Natural.

Dramatic voices

Emotionally expressive voices that understand context and feeling. They interpret the script and perform on their own, so you can start rehearsing immediately with no audio input required.

These voices analyze the text, punctuation, and context to deliver lines with natural emotion and pacing. They are the fastest to set up and work well for most rehearsal sessions.

Higher variance means the delivery can surprise you. Sometimes that is exactly what you want. Other times, you may need to regenerate a line to get a read that fits the scene.

When to use Dramatic voices

• You want to start rehearsing quickly without extra setup
• You want emotionally expressive delivery that responds to the text
• You are okay with occasional unexpected reads (you can always regenerate)
• You want to add speech instructions per line for overall scene context

Natural voices

Clear, even voices that read exactly what is on the page. Lower variance and a steady, predictable delivery every time. If you want a clean, neutral read with precise pacing, this is the type to use.

You cannot direct Natural voices. They ignore speech instructions and do not shape emotion, so you get the same clean, low-emotion read every time. That consistency is the point, it is what makes them great for self tapes, but if you need an emotional performance, use a Dramatic voice or the Voice changer instead.

When to use Natural voices

• You need reliable, predictable delivery for every line
• You are working on a scene where pacing and tone must be exact
• You are preparing a self tape and need consistent reads across multiple takes

Voice changer

Turn your own performance into an AI voice. It is the same idea as Record It Myself, with an AI voice laid on top. Record or upload a read of the line, with your own timing, emotion, and emphasis, and the voice changer keeps that delivery while swapping in the AI voice. You direct it by how you perform it, so it is a great fit for self tapes when you want a different voice but full control of the read.

When to use Voice changer

• You want to control the exact delivery yourself, line by line
• You have a reference read whose timing and emotion you want to keep
• You want a specific AI voice speaking your own performance

Cheap voices

Lower-cost voices that are cheaper to run than the ones we offer today. They are built for drafting and long rehearsal sessions where budget matters more than nuance. The Cheap filter is already in the picker but stays empty until these voices arrive. They are coming soon.

When to use Cheap voices

• You are drafting a scene and do not need the highest quality yet
• You are running long sessions and want to keep costs down
• You are happy to trade some nuance for a lower per-minute rate

Record It Myself

Use your own audio as a character's voice instead of a generated one. Record each line in the app, or upload any audio file you already have — a scene partner's read, a coach's take, or a recording you captured elsewhere. Whatever you provide plays back as-is, with no AI on top.

When to use Record It Myself

• You have recordings from a scene partner or coach
• You want to rehearse against a specific read you have already captured
• You need the most realistic version of your scene for self tape prep

Step 1: Assign the voice type

Open the voice picker for a character and choose Record it myself at the top of the list. This character no longer uses an AI voice, you'll record each of their lines yourself instead.

Step 2: Add audio to a line

On any line for that character, tap Add audio, then choose Record audio to speak the line yourself, or Upload a file if you already have a recording.

• Recording - tap Start Recording and read the line aloud. You have up to 4 minutes, with a warning at the 3-minute mark, then tap Stop Recording.
• Uploading - pick an audio file up to 5 minutes long.

Step 3: Trim and save

Fine-tune the start and end of your take, then tap Save. The line now plays back your recording instead of a generated voice.

Step 4: Change or remove

Once a line has a recording, use Change audio to record or upload a new take, Trim to adjust it again, or Remove to clear it. Each line holds one take at a time, recording a new one replaces the old take.

Punctuation

Punctuation matters for all voice types. Without it, the voice may guess wrong.

Bad:

Better:

Ellipses create pauses. Exclamation marks add energy. Question marks shift intonation. The more clearly your text signals the rhythm, the better the delivery.

Speech instructions

Speech instructions let you set the overall context for how a line should be delivered. This works with Dramatic voices.

Open the line settings (...) on any dialogue line and type a direction in the Speech instructions field.

Examples:

• "She just found out she got the part"
• "Cold and distant, holding back tears"
• "Late night, exhausted but relieved"
• "Rushed, overlapping with the previous line"

Speech instructions describe the scene or mood rather than marking up the text.

Troubleshooting: voice & audio