ElevenLabs
Expressive AI voices, cloning, dubbing and voice agents on one credit pool
Overview
ElevenLabs makes AI speech models and the products built on them: text to speech, voice cloning, dubbing, music and sound effects in ElevenCreative, conversational voice agents in ElevenAgents, and an API for developers. Its newest models, Eleven v4 and the low-latency v4 Turbo, arrived in September 2026 with support for more than 90 languages.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 14 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Text to speech, speech to text, sound effects, voice design, music and image
- 3 projects in Studio
- Commercial licence
- Instant voice cloning
- 20 projects in Studio
- Dubbing Studio and music commercial use
- Everything in Starter
- Professional voice cloning
- Additional credits
- Everything in Creator
- 44.1kHz PCM audio output via API
- 192kbps audio
- Everything in Pro
- 3 workspace seats
- Team collaboration
- 3 professional voice clones
- Everything in Scale
- Low-latency TTS as low as 5c a minute
- 10 professional voice clones
- 10 workspace seats
Use cases
What you can produce with ElevenLabs
- Narration and voiceover in 90+ languages from Eleven v4, directed with inline audio tags
- Instant voice clones from about 10 seconds of audio, or professional clones on Creator and above
- Dubbed video that keeps the original speaker's voice, via Dubbing v2
- Conversational voice agents on ElevenAgents, using the low-latency v4 Turbo model
- Transcripts from Scribe v2, in batch or real time
- Music tracks and sound effects generated from text prompts
- Per-character API access for building voice into your own product
ElevenLabs remains the reference point for synthetic voice. Eleven v4 adds finer control through audio tags and natural-language direction, v4 Turbo targets live agents at about 100 milliseconds of median inference by the company's measure, and one credit pool covers speech, cloning, dubbing and music. Unused paid credits now roll over for two months, which softens the old complaint about losing them.
The costs to watch are volume, since every generation spends credits, and governance for cloned voices. For audiobooks, voiceover, localisation and voice agents it is the default shortlist entry; heavy API users should compare the per-character rates against a subscription.
What we found
ElevenLabs is a voice AI company with a London headquarters, run through Eleven Labs Inc., that raised its first round in 2022 alongside its first speech model. It now sells three things from one account: ElevenCreative for text to speech, voice cloning, dubbing, music and sound effects; ElevenAgents for conversational voice agents; and ElevenAPI for developers. The newest models are Eleven v4 and Eleven v4 Turbo, released on 28 September 2026 and covering more than 90 languages. v4 is the expressive model, steered by inline audio tags such as [laughs] or plain-language direction, and Turbo is the agent model, with a median inference latency of about 100 milliseconds by the company's measurement. The preference and latency comparisons in the launch post are ElevenLabs' own tests against named competitors, not third-party results.
Subscriptions share one monthly credit pool across every product. The free plan gives 10,000 credits a month with no commercial licence; Starter costs $6 a month for 30,000 credits and adds commercial use and instant cloning; Creator is $22 for 121,000 credits with professional cloning; Pro is $99 for 600,000; Scale is $299 with three seats; Business is $990 with ten seats. Annual billing charges ten months for twelve. Unused paid credits roll over for up to two months, capped at twice the monthly quota, but downgrading or cancelling forfeits them. API buyers can instead pay per use: $0.08 per thousand characters for v4 and $0.04 for v4 Turbo.
The business has grown fast. A $500 million Series D led by Sequoia in February 2026 valued the company at $11 billion; by May it reported passing $500 million in annual recurring revenue; and on 30 September a $300 million employee tender, a sale of existing shares rather than new funding, priced it at $22 billion. Enterprise now makes up 55% of revenue, by the company's account. The things to manage are cost at volume, since every generation spends credits, and consent and misuse controls for any cloned voice.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
Notable facts
- ElevenLabs can clone a voice with as little as one minute of sample audio — a process that previously required hours of professional studio recording.
- The company grew from zero to 1 million users in its first year, driven almost entirely by word-of-mouth from podcasters and content creators.
- ElevenLabs' Eleven Multilingual v2 model was the first commercial TTS model to achieve near-human scores on the MUSHRA audio quality benchmark across 29 languages.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including ElevenLabs's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with ElevenLabs unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to ElevenLabs directly →
Spotted something out of date? Suggest an update →
More in Video & Audio
