ElevenLabs
by ElevenLabs
Getting-started guide
Turn text into remarkably natural speech — and, with permission, clone a specific voice — across dozens of languages.
Current prices: Free, from $5 equivalent, Creator $18.33, Pro $82.50) · See all plans
Last reviewed 2026-10-14 · for ElevenLabs
You paste text, choose a voice — a stock one, a community one, or one you've cloned from a sample — and it generates the audio. You listen, adjust the text or pick a different voice, and generate again until it sounds right, then keep the take you like.
ElevenLabs is an AI voice platform. Its core is text-to-speech that sounds convincingly human; it also does voice cloning (making a synthetic version of a specific voice from a sample), dubbing, and sound effects. It's used for audiobooks, video narration, podcasts, and voice agents. You mostly drive it by typing the text you want spoken and picking a voice.
The free tier gives a monthly credit allowance and lets you try text-to-speech and cloning — but its output is non-commercial and must credit ElevenLabs. Commercial use starts on the cheapest paid plan, Starter (official FAQ, 2026-10-14); Creator, Pro and Scale add more credits and features.
If you're going to clone a voice, sort out consent first — this is the single most important thing about the platform (below).
Credits are metered by characters of text and reset monthly; unused credits generally don't carry over.
When you upload a voice sample or generate audio, ElevenLabs keeps a broad, long-running license to use your voice data and content to improve its models. It commits not to commercialize your voice on its own without permission, but the training use is part of the deal — factor that in before uploading a sensitive or valuable voice.
On the free tier, output is non-commercial and must attribute ElevenLabs; commercial rights begin on the paid plans (and persist for audio you already generated, even if you later cancel). Don't build a paid project on free-tier audio.
Welcome to the show. Today we are talking about something small that changed everything: the humble paperclip.
Why this one: The fastest way to judge the quality — pick a built-in voice, paste text, generate, and listen. No cloning, no consent questions.
Tweak: Try the same line in two different voices to hear how much the voice choice matters.
[Record or upload a clean sample of your own voice reading a paragraph, then generate this line in your cloned voice:] Thanks for listening - I recorded this without touching a microphone twice.
Why this one: Cloning your own voice is the standout feature and the one clone you can make with zero consent worries.
Tweak: Use a quiet room and a minute or two of clear audio for a noticeably better clone.
Translate and speak this in Spanish, keeping a warm, friendly tone: "Hi everyone, thanks so much for being here today."
Why this one: Shows the multilingual side that makes ElevenLabs useful beyond English narration.
Tweak: Swap in a language your audience actually uses.
Want the full picture first? See the ElevenLabs overview for pricing, the free tier and related tools.
