Voice model · ElevenLabs v4
ElevenLabs is an AI audio company, and Eleven v4 is its most expressive voice model. On Deepnia, it turns up to 5,000 characters into a voice-over with a voice from the ElevenLabs catalogue, in one of the 85 languages in the menu (including French, English, Spanish, Arabic, Brazilian Portuguese, Swahili, Hausa, Wolof, Lingala and Japanese) or with auto-detect. Tags such as [whispers] or [laughs] direct the delivery, and three stability levels (Creative, Natural, Robust) set how expressive the voice is.
The voice generator opens with ElevenLabs v4 already selected. The exact cost shows before every creation; you sign up when you generate.
Put [whispers], [laughs], [sighs], [excited], [sarcastic], [crying], [shouting], [pause], [long pause] in your script: ElevenLabs v4 performs them as directions and never reads them out. The “Add tags with AI” button places them for you without rewriting your text.
Choose the language of your text in a searchable menu grouped by region: French, English, Spanish, Arabic, Brazilian Portuguese, Swahili, Hausa, Wolof, Lingala and Japanese, among others. With auto-detect, the language is inferred from the text.
Pick a voice from Deepnia’s ElevenLabs catalogue: search by name or style, then play its French or English preview before you generate.
Under “Advanced settings”, choose Creative, Natural (the default) or Robust. According to ElevenLabs, lower stability widens the emotional range, while higher stability makes the voice steadier.
Each generation delivers one WAV file (24 kHz, mono), an uncompressed format that drops easily into a video, a podcast or a presentation.
The “Try ElevenLabs v4” button opens Deepnia’s voice generator with the model already selected.
Search the ElevenLabs catalogue, play a voice’s preview, then choose the language of the text or “Auto-detect”.
Write up to 5,000 characters, then add tags such as [whispers] or [pause], or let AI place them.
Under “Advanced settings”, choose Creative, Natural or Robust, check the cost shown, then click “Generate voice”.
Copy a prompt, open the tool and adapt it to your project.
[excited] Every morning starts with a promise: coffee roasted the day before, just for you. [pause] Meet Aurora, our new blend, in store today.
Module two: simple bookkeeping for your business. First, we will sort your invoices; then we will track your expenses week by week.
That morning, the market woke long before sunrise. [whispers] Nobody had noticed it yet… [long pause] but the letter had arrived.
Bienvenue dans votre nouveau programme d’entraînement. Indiquez votre objectif : nous préparons votre première semaine de séances en moins d’une minute.
Heute stellen wir Ihnen unsere neue Smartwatch vor: leichter, schneller und mit einem Akku, der eine ganze Woche hält.
Welcome, everyone, to this special day! [laughs] Please take your seats; the ceremony begins in a few minutes. The couple thanks you for being here to share this moment.
Deepnia’s other voice and music models, to pick the right one for your project.
| Model | Inputs | Settings | Best for |
|---|---|---|---|
| ElevenLabs v4This model | Text | Language, Stability, Expression tags | ElevenLabs voices, directed with tags, in 85 languages. |
| ElevenLabs v4 Turbo | Text | Language, Stability, Expression tags | The same voices and languages, in a fast version. |
| Gemini Flash TTS | Text | Expression tags | A Google voice steered by tags, with the language detected automatically. |
| MiniMax Speech 2.6 HD | Text, Authorized personal voice | Language, Speed | A cloned voice, with adjustable speed and 23 languages + Auto. |
The menu offers auto-detect plus 85 languages, including French, English, Spanish, Arabic, Brazilian Portuguese, Swahili, Hausa, Wolof, Lingala and Japanese. It is searchable and grouped by region.
They direct the delivery: [whispers], [laughs], [sighs], [excited], [sarcastic], [crying], [shouting], [pause], [long pause]. ElevenLabs v4 performs them without reading them out. The “Add tags with AI” button can place them in your text.
For an ad, a warm or upbeat voice on Creative; for training, a calm voice on Robust; for an audiobook, a deep narrator voice on Natural, with a few tags. Play the previews to find the timbre that suits your audience.
It sets how expressive the voice is, in three levels: Creative, Natural (the default) and Robust. According to ElevenLabs, lower stability widens the emotional range, while higher stability makes the voice steadier.
Turbo is the fast version of ElevenLabs v4: the same catalogue voices, the same languages and the same tags.
Up to 5,000 characters per generation, delivered as one WAV file (24 kHz, mono). For a longer text, generate it in several parts and join them in your editor.
Yes: translate your script, then generate each version with its language picked in the menu, which offers 85 languages plus auto-detect.
The voice generator opens with ElevenLabs v4 already selected. The exact cost shows before every creation; you sign up when you generate.
