ElevenLabs
A voice AI platform for turning scripts into narration, plus dubbing and transcription.
Tools for turning scripts into speech, cleaning up recordings and moving audio between languages. Compare credit allowances and commercial licences before you publish.
Audio and voice AI tools turn written text into spoken audio, and handle transcription, dubbing and voice cloning alongside it. Usage is usually metered in credits rather than minutes, and the commercial licence you need in order to publish sits on the paid tiers — the free tier of the tool we profiled does not include it.
1 tool
Clear filtersA voice AI platform for turning scripts into narration, plus dubbing and transcription.
Two things decide whether a generated voice track is usable, and neither appears on a feature list. The first is how the voice handles numbers, names and punctuation: times, prices and unfamiliar surnames are where a track that sounded fine in a demo needs a manual fix. The second is pacing — whether the reading has the rhythm of someone explaining something, or the flat cadence of text being processed.
The practical constraint is the credit system. Credits are not minutes, and how far one goes depends on the feature and the audio quality you select, so a plan headline translates poorly into output. Estimate against a real script rather than the number on the pricing page.
Voice cloning deserves separate thought. The feature exists and is straightforward to use, which is exactly why the consent question matters: cloning a real person's voice requires their permission, and platform terms and local law both apply on top of that.
Read the ElevenLabs profile for our current feature, pricing and source notes. For a project that also includes footage, browse the video tools category. For script preparation, visit writing tools. ElevenLabs is currently the only profile in this category; we do not present it as the winner of a comparison.
Our editorial judgement about what matters in this category, based on the plan structures we documented while building these profiles.
Only on a plan that includes a commercial licence. On the tool we profiled, that starts at the paid entry tier and is absent from the free allowance. Decide where the audio will be published before choosing a plan, because that decision sets your minimum tier.
Credits are the usage unit used instead of minutes of audio. How far a credit goes depends on which feature you use and the audio quality you select, so the same credit balance produces different amounts of output depending on your settings.
The feature exists on paid plans, but cloning a person's voice requires their consent, and both the platform's terms and your local law apply. Treat it as a rule of the work rather than a setting to switch on, especially for identifiable real voices.
Not the right category? Browse every tool in the directory, put two options side by side, or read how we check and correct these profiles.