ElevenLabs v3
ElevenLabs's latest TTS — 70+ languages, 20 named voices (Aria, Roger, Sarah, ... Bill), inline audio tags ([laughs], [whispers], [excited]), and word-level timestamps for subtitling. Broadcast-quality output at $0.10 per 1,000 characters.
Model Specs
- Released
- Jun 2025
- Voices
- 20
- Languages
- 29
- Max characters
- 5K
- Modalities
- textaudio
About this model
ElevenLabs v3 (Eleven v3) is ElevenLabs's latest text-to-speech model and the most feature-complete TTS option on Renas AI. Three capabilities make it distinctive: (1) **70+ languages** including English, Mandarin, Hindi, Arabic, Spanish, French, German, Japanese, Korean, and dozens more — the broadest language coverage in any Renas TTS model; (2) **inline audio tags** like `[laughs]`, `[whispers]`, `[excited]`, `[sad]` for fine-grained emotional direction within the script; (3) **word-level timestamps** for subtitling and lip-sync workflows where you need to know exactly when each word is spoken.
The model offers 20 named voices: Aria, Roger, Sarah, Laura, Charlie, George, Callum, River, Liam, Charlotte, Alice, Matilda, Will, Jessica, Eric, Chris, Brian, Daniel, Lily, Bill — with Rachel as the default. Output formats include MP3, PCM, µ-law, A-law, and Opus codecs at various sample rates and bitrates (default mp3_44100_128). Pricing is $0.10 per 1,000 characters — same as Multilingual v2 — with no seat licenses, subscriptions, or minimums. ElevenLabs's voice cloning (Instant and Professional tiers) is referenced but typically accessed at the ElevenLabs platform level rather than via the fal.ai integration.
Reach for ElevenLabs v3 when (a) you need broad multilingual TTS coverage (70+ languages), (b) your content benefits from inline emotional direction via audio tags, (c) you're producing subtitled or lip-synced content where word timestamps matter, or (d) you want the latest ElevenLabs voice quality. For cost-sensitive workflows, Kokoro ($0.02) or Dia ($0.04); for established workflows on the older multilingual v2, that's still available on Renas at the same price.
Key Strengths
70+ language coverage
Broadest multilingual support on Renas TTS — English, Mandarin, Hindi, Arabic, Spanish, French, German, Japanese, Korean, and dozens more. Useful for international content workflows, multilingual podcasts, and global brand audio.
20 named voices
Hand-curated voice library: Aria, Roger, Sarah, Laura, Charlie, George, Callum, River, Liam, Charlotte, Alice, Matilda, Will, Jessica, Eric, Chris, Brian, Daniel, Lily, Bill. Default is Rachel. Pick voices by name and character — easier than navigating large unnamed libraries.
Inline audio tags for emotion
Direct emotional direction within the script via tags like `[laughs]`, `[whispers]`, `[excited]`, `[sad]`. Combined with broad dynamic range, produces expressive output without separate parameter tuning.
Word-level timestamps
Output includes precise per-word timing — essential for subtitling workflows, lip-sync video production, and any context where you need to know exactly when each word is spoken.
Multiple output codecs
MP3, PCM, µ-law, A-law, and Opus formats at various sample rates and bitrates. Default mp3_44100_128 covers most use cases; specialized codecs available for telephony (µ-law/A-law) or low-bandwidth (Opus).
Contextual emotional understanding
Beyond explicit tags, v3 reads narrative cues from context — adjusting tone based on the surrounding text without requiring you to tag every emotional shift.
Voice synthesis capabilities
Available voices, languages, and expressive controls.
How it compares
ElevenLabs v3 is the most feature-rich TTS on Renas. Compare against alternatives based on language coverage, cost, and feature requirements.
| vs. Model | Verdict | Outcome |
|---|
Pros
- 70+ language coverage — broadest on Renas TTS
- 20 hand-curated named voices
- Inline audio tags ([laughs], [whispers], [excited], [sad]) for emotion
- Word-level timestamps for subtitling and lip-sync
- Multiple output codecs (MP3, PCM, µ-law, A-law, Opus)
- Contextual emotional understanding from narrative cues
- Broadcast-quality voice synthesis
Things to consider
- Premium pricing ($0.10/1K chars) — 2.5-5x more than Dia/Kokoro
- Max input characters not specified in fal.ai page
- Voice cloning happens at ElevenLabs platform level, not via fal.ai integration
- No multi-speaker dialogue tags (Dia has [S1]/[S2])
- Specific accent and voice variations within named voices not fully documented
Best use cases
Multilingual content production
Generate the same content in multiple languages — Spanish marketing video, Mandarin product descriptions, Arabic news content. 70+ languages lets you scale to global audiences from a single TTS model.
Subtitled and lip-synced video
Word-level timestamps enable precise subtitle generation that matches the audio. Useful for accessibility-compliant video, multilingual subtitles, or animation/lip-sync workflows.
Audio dramas and narrative content
Inline audio tags ([laughs], [whispers], [sighs]) give fine-grained emotional control for character-driven audio. Combine with 20 named voices for casts of characters in audiobooks, podcasts, or audio dramas.
Premium podcast production
Broadcast-quality output + 20 named voices + emotional control fits podcast workflows where production quality matters. Use Rachel as default narrator, switch to other voices for guests or characters.
International marketing campaigns
Generate brand voice content across all your target markets in a single workflow. Same script, different language outputs — useful for global ad campaigns and content marketing at scale.
Accessibility content
Audio descriptions for visual content, screen reader-friendly content alternatives, multilingual accessibility transcripts. 70+ language support combined with broadcast quality makes v3 a strong accessibility tool.
How to use it on Renas AI
- 1
Step 1
Open the AI Voice tool in TTS mode
Navigate to AI Voice in the Renas dashboard, then switch to Text-to-Speech mode. Pick ElevenLabs v3 from the model selector — it's marked as the latest ElevenLabs variant.
- 2
Step 2
Pick voice and language
Choose from 20 named voices (Aria, Roger, Sarah, etc.) — Rachel is default. The 70+ language support is automatic — write in your target language and v3 handles it.
- 3
Step 3
Add inline audio tags for emotion
Direct emotional delivery in your script: `Hello, [excited] welcome to the show!` or `[whispers] This is a secret. [laughs] Just kidding.` Tags like [laughs], [whispers], [excited], [sad] work inline.
- 4
Step 4
Generate, review, deploy
Output goes to your asset library in your selected codec (default MP3). For subtitling, capture word-level timestamps from the API response. For video lip-sync, use the timestamps to align audio with visual elements.
Pricing
Pricing on Renas AI
Pay-as-you-go credits, no API keys, no rate limits.
Frequently asked questions
Other ElevenLabs models
Other voice models on Renas AI
70+ language premium AI voice
Use ElevenLabs v3 with your Renas AI subscription credits — no API key, no setup, no per-seat fees.
Try ElevenLabs v3