AI generátor hlasu a platforma hlasových agentov s realistickou viacjazyčnou rečou a API
Zvukové nástroje AI try.elevenlabs.io
What is ElevenLabs
ElevenLabs is a speech synthesis software company founded in 2022 by Polish entrepreneurs Piotr Dąbkowski, a former Google machine learning engineer, and Mati Staniszewski, formerly of Palantir. It is based in London, with offices in New York, Warsaw and San Francisco. The "Eleven" in its name refers to 11 November, Poland's National Independence Day.
What it can do
According to the official documentation:
- Text to speech: the Eleven v4 model covers 90+ languages with a consistent voice when switching between them; v4 Turbo targets real-time use at a median latency of about 100 ms; Flash v2.5 costs less and supports 32 languages
- Speech to text: Scribe v2 transcribes 90+ languages with word-level timestamps and can separate up to 32 speakers; the real-time version has about 150 ms of latency
- Voice agents: deploy AI support agents and assistants that talk by voice or chat
- Voice cloning, plus music and sound effects generation
- APIs and SDKs for developers
Pricing
Official monthly prices:
- Free: no charge, 10,000 credits
- Starter: US$6, 30,000 credits
- Creator: US$22 (US$11 for the first month), 121,000 credits
- Pro: US$99, 600,000 credits
- Scale: US$299, 1,800,000 credits, 3 seats
- Business: US$990, 6,000,000 credits, 10 seats
Who it's for
- Audiobooks, video voiceovers and podcasts
- Teams that need multilingual voiceover and localisation
- Developers building voice support agents and assistants
Good to know
- Cloning someone else's voice requires their permission, and platforms restrict impersonation.
Similar tools
Compare Speechify and Moyin; for music, see Suno.