We read every ElevenLabs (v2 / v3) prompt guide so you don't have to.

What ElevenLabs actually rewards

We pulled this from ElevenLabs's official guidance and what works in production. The short version:

• →Match tags to voice character (don't put [shout] on a whisper-trained voice).
• →Use punctuation deliberately (ellipses for pauses, ALL CAPS for emphasis).
• →Experiment with stability 25-50% for emotional variety.
• →Use Voice Design with rich preview text matching desired tone.

Common mistakes we fix automatically

• Avoid
  Don't use <break> in v3 (not supported).
• Avoid
  Don't mix many tags per line (instability).

Frequently asked questions

Which versions of ElevenLabs (v2 / v3) does this support?

We support v2, v2.5, flash, multilingual, v3-alpha, v3-conversational. We apply the prompt patterns ElevenLabs recommends for each, so the rewrite is tuned to the version you're using.

Is my prompt stored or used for training?

No. Prompts are sent to the rewriter, scored, returned, and discarded. We don't train on them and we don't keep them around.

Do I need to know prompt engineering to use this?

Nope. That's the point. Paste what you have, click Rewrite, get back a version that follows ElevenLabs's official guidance.

What makes this different from ElevenLabs (v2 / v3)'s own "improve prompt" feature?

Built-in optimizers use the model's own preferences. Ours is built on ElevenLabs's official documentation and patterns that consistently produce better results in production. We keep rewrites inside the length window ElevenLabs (v2 / v3) responds best to.