Voice model
eroq Voice One — the expressive voice
eroq Voice One is the studio's flagship text-to-speech engine: natural, expressive reads from a curated roster of 24 voices (Aria, warm and intimate, and Orion, low and unhurried, among them), from voices you clone from your own audio, or from a Character's assigned voice. Delivery is driven by punctuation and two controls — speed and expressiveness — and the output is an MP3 you can drop anywhere.
打开工作室并自动填好内容。生成时再注册:免费送 40 积分,无需绑卡。
- Price
- 3 credits / 100 characters
- Voices
- Roster of 24 · your clones · Characters
- Languages
- 13
- Controls
- Speed 0.7–1.3 · expressiveness 0–1 · punctuation
- Output
- MP3
- Access
- Every account
Directing a read
The engine reads punctuation like an actor reads a script: commas breathe, an ellipsis hesitates, a dash cuts, a question lifts. Expressiveness at 0 is a steady narration; at 1 it is theatrical. Speed stays around 1 for dialogue and drops for audiobook-style narration. The Delivery guide in the voice studio lists the six techniques with examples you can insert.
Cloning and casting
- →Clone a voice from a clean recording — a minute of natural speech, no music, one speaker.
- →Assign a voice to a Character so every line they speak, in any scene, is the same voice.
- →Use `char:<id>` as the voice in the API to speak as a Character.
从这里开始
选一个:它会填进顶部的输入框,直接就能生成。
常见问题
How much does text-to-speech cost on eroq?
3 credits per started block of 100 characters on eroq Voice One, 2 on eroq Voice Turbo. A 1,000-character paragraph is 30 credits — about $0.30 at the entry pack rate.
Can I clone my own voice?
Yes, from an audio upload in the voice studio. Clones are private to your account and usable in the studio and the API.
Which languages are supported?
Thirteen, on both speech engines; the roster voices read English natively and the other twelve on top.
Is there a faster or cheaper option?
eroq Voice Turbo — lower latency, 2 credits per 100 characters, same roster and clones.
继续探索









