ElevenLabs

What it is

ElevenLabs turns text into speech that sounds like a person reading it, and clones voices from a sample so the speech sounds like a specific person. Around that sit speech to text, sound effects, voice design, music, image tools and Studio, where you assemble longer productions.

Dubbing Studio is the feature that changes what small teams can do: an existing recording translated and re-voiced into another language. On the free plan you get the generation tools with three Studio projects, and crucially no commercial license, which means free output is for trying things, not for publishing.

Who it's for (and who it isn't)

It fits people who need voice output regularly and cannot record it every time: course narration, product videos, audio versions of written content, ads in several languages. It also fits anyone whose own recording setup is the bottleneck, since a cloned voice can narrate a script you never spoke.

It is a weaker fit where the voice is the relationship. A podcast built on your presence, or content whose value is that a real person said it, loses something that quality does not replace. There is also a disclosure question worth settling before you publish: audiences increasingly notice synthetic narration, and the cost of being caught not saying so is higher than the cost of saying so.

Pricing

ElevenLabs is freemium and priced on monthly credits. The free plan gives 10,000 credits per month, access to text to speech, speech to text, sound effects, voice design, music and image tools, and 3 Studio projects, with no commercial license and no voice cloning. Starter is $6 per month for 30,000 credits and adds a commercial license, instant voice cloning and Dubbing Studio. Creator is $22 per month for 121,000 credits with professional voice cloning. Pro is $99 for 600,000 credits, Scale $299 for 1,800,000, Business $990 for 6,000,000, and Enterprise is custom.

The commercial license is the line that matters more than the credits. If the output goes anywhere public, the free plan is not an option, and Starter at $6 per month is the actual entry price.

Set it up

Official guide: ElevenCreative quickstart

1. Create an ElevenLabs account on the Free plan. It gives you 10,000 credits per month (credits are the monthly allowance that each piece of audio you generate uses up) and has no commercial license, so treat everything you make as practice. Have one paragraph of your real script ready. Your free account is now ready.
2. Open the Voice Library and search by language, accent, age and use case. Click the orb next to a voice name to hear a short preview, and listen to two or three voices. You now hear a preview each time you click, so you can pick one voice.
3. Open the Text to Speech playground and select your voice from the dropdown. Paste your paragraph into the text box (up to 5,000 characters) and click "Generate". You now have audio of your paragraph that you can play.
4. Play the audio while you read your script. Mark each sentence where the voice misreads your meaning, rewrite it and click "Generate" again. You now have a second version to compare with the first.
5. Download or share the version you prefer. The audio you chose is now saved or shared outside the playground.

When this works, you have an audio file of your own paragraph read by a voice you chose, and you can judge whether narration without booking recording time fits your work.

On Starter ($6/mo): 30,000 credits, instant voice cloning (a copy of a voice made from a recording) and a commercial license, which you need before you use the audio commercially. On Creator ($22/mo): 121,000 credits.

Strengths and limitations

The strength is quality per euro. Narration that would have meant a recording session and a voice actor now costs credits, and Dubbing Studio makes a second language a production step instead of a project.

The limitations are ethical and practical in equal measure. Cloning someone else's voice without clear permission is a line worth treating as absolute. The free tier has no commercial license, so it is a sandbox. Credits meter everything, and long-form projects consume them faster than people estimate. And synthetic delivery still flattens emphasis on complicated sentences, so scripts need to be written for the ear rather than handed over as-is.