You have to pay for everything Good professional voice cloning, difficult to distinguish from the original. But there are too many problems with incorrect accent or pronunciation. I have a subscription, which supposedly should have been enough for 2 hours of audio. In fact, the video was enough for 6 minutes. It will be cheaper to order the services of a professional announcer. Most use cases for this product fall through the cracks due to poor pricing plans. I need a plan with an unlimited number of generations or to somehow solve the problem with low-quality generations for which I pay money.

Charge not for result, but for experiments I started with 10000 free "characters". That was not enough for testing. I signed up with another account and continued testing. All quota ended in 20 mins. Again. But I finally found voice and settings that worded for me. I srarted "Start" plan with 30000 characters quota. I barely could record 2 minutes voiceover and all my MONGTH quota has ended. Elevenlabs state that 30000 characters equals 30 mins of audio. No, it is not. It is impossible to record from 1st try. EVERY run will eat your quota. Despite it is test run, not real product yet. So after I corrected all pronunciation errors, I ended up with nothing left. To get real 30 mins of record (which I really need) I have to buy $99 Pro plan (Instead of $5 plan as advertised). Generation quality is great, Elevenlabs made a great product. But I think this is not fair to resell computer time instead of real result.

It is a promising platform for… It is a promising platform for generating AI audios, and maybe in a few years it will become useful. However, right now they exercise false advertising practices. They steal $100 per month from customers, promising real human-like voices in their audios. Then, when you try to publish their audios on Audible, Amazon rejects them saying that it is blatantly AI. When I reached out to ElevenLabs customer support, they said that they refuse to help with this and they have no idea why their generated voices sound like AI instead of human. All-in-all they basically stole money from me and gave false advertising of their product.

WHAT REALISTIC VOICES!!!! Never have I seen a voice cloning tool do anything so impossible in all my life! :). I swear, I’m not only listening to realistic sounding voices but voices that only needed like 1 minute of source material as opposed to 15 minutes to 2 hours. I SO can’t wait for your voice conversion feature to come out; good luck with that!

Excellent! The fact that there are such cheap prices with such astounding qualities of work is absolutely insane. A personal favourite feature of mine is the voice cloning in the speech labs. With even a small amount of HQ samples, the output is phenomenal!