Eleven v4 AI Voice Generator

Create expressive speech from a script. Direct one preset voice with audio tags for emotion and pacing, then keep every MP3 private.

44.1 kHz MP3
Preset voices
Audio tags
ElevenLabs · Eleven v4 docs ↗

Eleven v4 Voice Examples

Official ElevenLabs Eleven v4 mystery narration demonstration

Cold coffee

Prompt

They said she disappeared on a Tuesday. No note, no struggle, no trace. Just an open window and a cold cup of coffee. Nobody talked about it much after that. Nobody wanted to. But everybody knew.

Key Features of Eleven v4

Demonstration audio published by ElevenLabs. Audio tags, expressive delivery, a character read, and a Spanish script—each row is one voice. Prefill any row; this route generates a single preset voice, and nothing runs until you generate.

Audio tag direction

Direction sits in the script. This ElevenLabs demonstration breaks into laughter on the pigeon punchline, with tags such as [chuckles] and [laughing].

Official ElevenLabs Eleven v4 realism demonstration with laughs

Best for / Not for

Use Eleven v4 when a single preset voice should follow tags in the script.

Best for

Short ads, launch copy, and explainer voice audio

Scripts that mark emotion with inline audio tags

Reviewed preset voices for a publishable take

Lines written in the language you want heard

Not for

Voice cloning or an uploaded voice

More than one speaker in a single take

Live agents or low-latency conversation

Scripts longer than 5,000 characters in one request

How to Use Eleven v4

Three steps from script to a reviewable voice audio take.

1
1. Write the script
Enter the approved copy. Use commas and periods for pacing. Prefill from an Example or Key Feature if you want. Nothing runs until you generate.
Enter the approved copy. Use commas and periods for pacing. Prefill from an Example or Key Feature if you want. Nothing runs until you generate.
2
2. Set the voice and tags
The route defaults to Rachel and also offers Aria, Roger, Sarah, and George. Put a tag such as [excited] or [whispers] immediately before the line that should change.
The route defaults to Rachel and also offers Aria, Roger, Sarah, and George. Put a tag such as [excited] or [whispers] immediately before the line that should change.
3
3. Generate and review
Confirm the credit estimate, generate, then listen for timing, pronunciation, and delivery. Adjust the tags or the preset and run again before you download.
Confirm the credit estimate, generate, then listen for timing, pronunciation, and delivery. Adjust the tags or the preset and run again before you download.

Create with Eleven v4 ↑

Eleven v4 Pricing

From 6 credits per 1,000 characters on the published route. Every run uses credits, with the exact quote shown before submission.

Compare Eleven v4How Eleven v4 compares to other models

Tag-directed preset voice audio, a directed sound scene, and the previous Eleven voice route.

Eleven v4

This model

Tag-directed single-speaker voice audio

Private speech audio6 credits per 1,000 characters

Seed Audio 1.0 →ByteDance

AI video beds, short-form scenes, and timed multi-speaker dialogue

Private speech audio14 credits per take

Eleven v3 →ElevenLabs

Emotion-tagged short VO and character performance

Private speech audio7 credits per 1,000 characters

Attribute	This modelEleven v4	ByteDanceSeed Audio 1.0 →	ElevenLabsEleven v3 →
Best for	Tag-directed single-speaker voice audio	AI video beds, short-form scenes, and timed multi-speaker dialogue	Emotion-tagged short VO and character performance
Performance control	Inline audio tags and IPA in the script	Scene-style text direction in one brief	Inline audio tags ([laughs], [excited], …)
Voices	5 named presets (no cloning)	Model-selected delivery (no preset picker)	5 named presets (no cloning)
Mode	Single-speaker speech	Text-directed speech and audio scenes	Single-speaker speech
Output	Private speech audio	Private speech audio	Private speech audio
Credits	6 credits per 1,000 characters	14 credits per take	7 credits per 1,000 characters

Stay on Eleven v4 for a single preset voice with audio tags. Switch to Seed Audio 1.0 for a directed sound scene, or to Eleven v3 for the previous Eleven voice route.

ElevenLabs Eleven v4

Eleven v4 FAQs

What is Eleven v4?

Eleven v4 is ElevenLabs’ expressive text-to-speech model. On MixVio you write a script, add audio tags when a line should change, choose a reviewed voice preset, and generate a private audio take. Credits are shown before you run.

What is Eleven v4 best for?

Use it for launch copy, explainers, and character-style reads where emotion tags and delivery matter as much as the words. This published route is a single-speaker voiceover, generated and reviewed before you download.

How do audio tags work?

Put direction in square brackets inside the script, for example [excited], [whispers], [laughs], [warm], or [softly], right before the line that should change. Tags are part of the script, not a separate control. Use commas and periods for pacing, and avoid stacking conflicting tags on the same line.

Which voices can I choose?

The public Eleven v4 route exposes five presets: Rachel, Aria, Roger, Sarah, and George. Rachel is the default. Voice cloning, Professional Voice Clones, and uploaded voices are not on this route.

Can I generate more than one speaker in a single take?

No. Each Eleven v4 generation uses one preset voice. For a second role, generate another take and edit the files together. Seed Audio 1.0 is the route for a directed sound scene from one brief.

How long can one Eleven v4 script be?

This published route accepts up to 5,000 characters per generation, matching the fal text limit. Split a longer script into separate takes. MixVio does not stitch context across requests on this route.

Can I use SSML or a language menu?

Write pauses and delivery as audio tags, not SSML. There is no language picker. The model follows the language of the script. Put a custom pronunciation between slashes when a name or acronym needs one, then review the take.

What does Eleven v4 output, and how many credits does it use?

Each run returns one private MP3. Pricing is 6 credits per started 1,000 characters, up to 30 credits at 5,000 characters. MixVio shows the exact reservation before submission. Failed runs cost no credits.

Is Eleven v4 free?

Yes on the Free plan when you have credits. Free voiceovers include an audible MixVio disclosure when that delivery rule is on. Paid-plan takes do not add that disclosure.

How does Eleven v4 compare with Eleven v3?

Both are single-speaker preset voiceovers with audio tags. Eleven v4 is the newer route at 6 credits per started 1,000 characters. Eleven v3 stays available at 7 credits per started 1,000 characters. Neither route clones a voice or runs a live agent.

Can I use Eleven v4 outputs commercially?

Paid-plan outputs may be used commercially, subject to the Terms of Service. You must have rights to the script. Free-plan outputs are for evaluation and are not for commercial publishing. Review pronunciation before you publish.

Direct expressive Eleven v4 voice audio.
Private. Credits shown first.

Write the script, add audio tags, pick a preset, then generate with Eleven v4 selected.