ElevenLabs
ElevenLabs: AI voiceovers, voice cloning and transcription
Create AI voiceovers, clone your voice and transcribe recordings with ElevenLabs. Free credits are available; commercial voiceovers require a paid plan.
Free plan availableAccount required

- Access
- Free and paid plans
- Example plan
- USD 6 / monthStarter — monthly renewal
- Available on
- Web, API
A voice platform for narration and interactive speech
ElevenLabs brings script-to-speech, custom voices and transcription into one audio platform. For video narration, audiobooks and character dialogue, you can choose a library voice, describe a new voice with Voice Design, or create a clone. The browser workspace suits creators producing spoken content; the API adds speech generation and transcription to applications.
The two cloning options have different requirements. Starter includes Instant Voice Cloning, for which ElevenLabs recommends one to two minutes of clear, consistent audio and requires rights and consent. Creator adds Professional Voice Cloning: you record your own voice, complete verification and wait for training. The professional guide recommends at least thirty minutes of speech, preferably two to three hours. Someone else must create and verify their own professional clone before sharing it with you.
Eleven v4 supports over ninety languages. Audio tags direct delivery, while Stability and Similarity adjust consistency and adherence to the reference voice. It has no SSML or Speed and Style sliders. Cross-language generation adopts the target language’s accent; professional-clone access is still rolling out. Older models have different language coverage and controls.
Scribe turns uploaded audio or video into timestamped text, with automatic language detection and speaker labels. Batch transcription supports over ninety languages and up to thirty-two speakers. The transcript editor lets you correct text, adjust timing and reassign speakers, then export a transcript or subtitles. Scribe v2 Realtime is a separate option for streaming applications.
For published commercial voiceovers, the plan used to generate the audio matters. Free output is noncommercial; upgrading later does not change its licence. Paid-generated content can retain commercial permission after cancellation, subject to applicable terms, rights and the exclusion for Beta Services. Voice clones remain in your ElevenLabs account rather than becoming downloadable models. Cloud processing and data-use settings also matter when working with recordings you want to keep private.
Best for
- Creators making narration and character dialogue
- Publishers producing multilingual spoken content
- Developers adding speech and transcription to applications
Limitations
Speech, transcription and other creative features share credits, so using one reduces what remains for the others. Free-generated voice output cannot be used commercially. Professional cloning requires the voice owner’s verification. v4 controls and accent handling differ from older models, and cloned voices are not portable model files.
ElevenLabs voice and audio features
Multilingual text-to-speech
Turn scripts into spoken audio with model-specific language coverage and voice controls.
Directed vocal delivery
Eleven v4 accepts audio tags for expressive delivery, alongside Stability and Similarity settings.
Custom voices
Instant cloning uses short recordings with rights and consent. Professional cloning requires your own voice, longer recordings, verification and training; it starts on Creator.
Voice library and design
Choose library voices or describe a new voice with Voice Design. Free API access excludes Voice Library voices.
Scribe transcription
Transcribe audio or video, identify speakers and edit timestamped text. Export plain text, JSON, HTML, SRT or VTT; browser uploads accept files up to 3 GB.
Application-ready audio
Use the speech API for generated audio and streaming, with model-, codec- and plan-specific output options.
ElevenLabs pricing
Free plan available. Subscription plans. Visit ElevenLabs for full plan details and current offers.
Example plans.
Starter — monthly renewal
USD 6 / month
ElevenCreative
Plan details
30,000 shared credits/month; Instant Voice Cloning; standard renewal price, excluding taxes.
Creator — monthly renewal
USD 22 / month
ElevenCreative
Plan details
121,000 shared credits/month; Professional Voice Cloning; standard renewal price, excluding taxes.
Pro — monthly
USD 99 / month
ElevenCreative
Plan details
600,000 shared credits/month; 44.1 kHz PCM through the API.
Technical specifications
Voice generation
| Feature | ElevenLabs |
|---|---|
| Speech languages | Eleven v4 supports 90+ speech languages. Multilingual v2 supports 29 and Flash v2.5 supports 32; coverage depends on the chosen model.Model-specific language coverage |
| Voice and pronunciation controls | Eleven v4 uses Stability and Similarity settings plus audio tags for delivery. Style and Speed sliders and SSML are unavailable in v4; other models have different controls.Text-to-speech model controls |
| Voice cloning | Starter includes instant cloning with rights and consent. Creator adds professional cloning of your own verified voice; another speaker must create and verify their clone before sharing it. v4 access is rolling out.Starter / Creator+; model availability applies |
| Audio exports and streaming | MP3 speech output, with API formats including PCM, Opus and telephony codecs. Higher-quality outputs have plan requirements; Pro lists 44.1 kHz PCM through the API.Codec and quality depend on plan |
| Commercial-use terms | Free voice output is noncommercial and requires title attribution when shared. Paid-created output carries commercial permission under the terms, except Beta Services; cancelling later does not remove that permission.Licence follows the plan used to generate the content |
| Output sample rate (Hz) | 44100Pro: 44.1 kHz PCM output |
| API access | SupportedSpeech generation and Scribe transcription APIs |
Transcription and meetings
| Feature | ElevenLabs |
|---|---|
| File and meeting capture | Upload audio or video in the browser or integrate Scribe through the API. Realtime transcription uses a separate Scribe v2 Realtime model.File transcription and application streaming |
| Transcription languages | Scribe v2 supports 90+ transcription languages and automatic language detection. Language coverage differs from speech-generation models.Batch transcription |
| Speaker-label workflow | Batch transcription supports up to 32 speakers and word-level timestamps. The transcript editor lets users rename and reassign speakers.Speaker diarization and editable labels |
| Transcript exports | Plain text, JSON, HTML, SRT and VTT. The browser editor supports corrections, segment timing and subtitle editing.Transcript and subtitle exports |
| Recording and usage limits | Browser audio/video uploads support files up to 3 GB. Transcription uses shared credits; other product usage reduces the remaining allowance. Keyterm and experimental editing features may add charges.3 GB file limit; shared subscription allowance |
| API access | SupportedBatch and realtime are separate models |
ElevenLabs alternatives
These platforms approach voice production and audio editing in different ways.
ElevenLabs FAQs
Is ElevenLabs free?
Yes. The free plan includes 10,000 monthly credits for its creative tools. Free-generated voice audio is noncommercial. When shared, it requires elevenlabs.io or 11.ai in the title; Eleven Music has its own attribution rule.
Can I use ElevenLabs audio commercially?
Paid-generated voice audio includes commercial permission under the applicable terms and third-party rights, except for Beta Services. That permission can continue after cancellation. Audio made on the free plan stays noncommercial after an upgrade.
Which ElevenLabs plan includes voice cloning?
Starter includes Instant Voice Cloning from short recordings, with rights and consent. Creator and higher plans include Professional Voice Cloning. Professional clones must be created and verified by the voice owner, even when another person has permission to use their voice.
Can ElevenLabs transcribe audio and video?
Yes. Scribe supports audio and video uploads, speaker labels and word-level timestamps. The browser file limit is 3 GB. You can edit the transcript and export plain text, JSON, HTML, SRT or VTT. Transcription uses the shared credit balance.
How much does ElevenLabs cost?
Standard monthly ElevenCreative prices start at USD 6 for Starter. Creator is USD 22 and Pro is USD 99, excluding taxes. First-month discounts are introductory prices. Annual billing and larger team plans have separate commitments and allowances.
How do ElevenLabs credits work?
The monthly credit pool is shared across creative products. Consumption depends on the task and model, so credits do not represent a fixed number of finished minutes. Generation requests can cost credits even if you do not download the result. Paid rollover requires an active subscription; cancelling or downgrading forfeits unused paid credits at the end of the cycle.
Can I download my ElevenLabs voice clone?
You can download generated speech, but the clone itself cannot be exported as a standalone model. It stays in your account. Other applications can use the voice through the ElevenLabs API; keeping your source recordings lets you create another clone later.
Does ElevenLabs use my recordings to train its models?
ElevenLabs can use submitted data to improve its models. You can disable the model-improvement setting under Terms and privacy → Data use; the opt-out applies to new data you submit. Enterprise data is excluded by default, except as necessary to provide the service. Audio processing remains cloud-based.
Explore related AI tools
Is this your product? Claim this page to update your listing.