YouTube2Text

Create Custom Realistic AI Avatars That Look & Sound 100% Like You (Full Workflow) — Transcript

by ElevenLabs · 2,333 words · 314 segments · language en · Watch on YouTube

Full transcript

  1. 0:00Imagine you could create a custom AI
  2. 0:01avatar that looks so real that people
  3. 0:03didn't actually know it was you. Well,
  4. 0:05you can. What you just saw was a digital
  5. 0:07clone of myself that looks and sounds
  6. 0:08100% like me. And if you didn't notice,
  7. 0:10it's because I created it using the 11
  8. 0:12Labs creative platform. There's a lot of
  9. 0:14different AI tools out there nowadays,
  10. 0:15but when it comes to creating AI avatars
  11. 0:17and cloning yourself, the quality is the
  12. 0:18most important thing. And thankfully, 11
  13. 0:20Labs provides all of the best AI image
  14. 0:22and video models out there and voice
  15. 0:24clone technology, allowing you to make
  16. 0:25the most realistic AI avatar clones out
  17. 0:27there.
  18. 0:27>> What do you think? Believe it or not,
  19. 0:30I'm [music] also an avatar from 11 Labs.
  20. 0:33>> And now, if I put myself side by side
  21. 0:34with my AI avatar and I lip-sync the
  22. 0:36audio, can you really spot the
  23. 0:38difference? And the cool thing is is
  24. 0:39that you can even change the clothes and
  25. 0:41put yourself in any environment. So, I
  26. 0:42decided to make myself look much smarter
  27. 0:44than I actually am. Hey everyone, my
  28. 0:46name is Alec and today I'm about to show
  29. 0:48you how to get the most realistic clone
  30. 0:49of yourself so you can simply type out
  31. 0:51what you want on your keyboard and
  32. 0:52create your video without having to go
  33. 0:54through the painful process of
  34. 0:56recording. So every take from now on is
  35. 0:58going to be perfect. Let's begin. Now
  36. 0:59with 11 Labs, you can create avatars
  37. 1:01that look 100% like you or you can make
  38. 1:03entirely AI generated avatars and you
  39. 1:05can even do a mix so you can put
  40. 1:07yourself in any [music] environment. But
  41. 1:09first, let's take a look at how to clone
  42. 1:11your voice. So you can create voiceovers
  43. 1:12that sound like you without actually
  44. 1:13having to record them. To clone your
  45. 1:15voice, you need to click on voices. And
  46. 1:17then on the right, we want to click on
  47. 1:18create a voice clone. And here we have
  48. 1:20multiple options to clone our voice.
  49. 1:22First of all, we got instant voice
  50. 1:23clone, which allows us to clone our
  51. 1:25voice with a minimum of 10 seconds of
  52. 1:27audio. And then we have professional
  53. 1:29voice clone, which allows us to create a
  54. 1:31super realistic digital replica of our
  55. 1:33voice. And here [music] we need to
  56. 1:34submit a minimum of 30 minutes of audio.
  57. 1:36And now, the higher the quality of the
  58. 1:38audio you submit to your voice clones,
  59. 1:40the better the result of the voice clone
  60. 1:42will be. And so, if we click on
  61. 1:43professional voice clone, as you can see
  62. 1:45here, you simply want to click create a
  63. 1:47new clone. And now go through the
  64. 1:48process of uploading your samples. You
  65. 1:50can even record directly into 11 Labs
  66. 1:52and [music] then name your voice clone.
  67. 1:54And if you want an in-depth guide to
  68. 1:55creating the highest quality voice clone
  69. 1:57possible, we'll put a link down in the
  70. 1:58description below to a full 10-minute
  71. 2:00[music] guide that we've put together
  72. 2:01for you. Once you have your voice clone,
  73. 2:02we can then click on text to speech. And
  74. 2:04here we can type out anything that we
  75. 2:06want our AI avatar to say. And so here
  76. 2:09I've got a default script talking about
  77. 2:11coffee. But to use our voice clone, we
  78. 2:13can then click on voices. And here we
  79. 2:15can go ahead and search for our voice
  80. 2:16clone. But as you can see here, I've got
  81. 2:18my own professional voice clone that
  82. 2:19I've named Alex PVC. And so if I click
  83. 2:22on it, [music] I can now generate this
  84. 2:24text right here that I've typed in, and
  85. 2:26it will sound just like me. So, let's
  86. 2:28take a listen. Coffee is one of those
  87. 2:29things most of us consume every day
  88. 2:31without really thinking about it. And
  89. 2:33so, as you can see, I've now got a
  90. 2:35script that sounds just like me, and all
  91. 2:37I've done is typed it out. And so, I
  92. 2:39could go in and replace this with
  93. 2:41anything I like. And so, if I said,
  94. 2:42"Hello YouTube, please subscribe." We
  95. 2:44can then click on generate. Hello
  96. 2:46YouTube, please subscribe. And as you
  97. 2:48can see, we've now got Hello YouTube,
  98. 2:50please subscribe in my voice. And so now
  99. 2:53that you have your voice clone and your
  100. 2:55voice over, we actually need to make the
  101. 2:57AI avatar. And to do this, we can head
  102. 2:59to the image and video tool on the left
  103. 3:01right here. And with inside image and
  104. 3:03video, there's a bunch of different ways
  105. 3:04that we can use and create AI avatars.
  106. 3:06We can use default avatars. We can
  107. 3:09create our own from scratch, but we
  108. 3:11could also upload pictures of oursel and
  109. 3:13even customize those avatars to get the
  110. 3:15exact result that we want. And so to
  111. 3:16begin, let's start with the default
  112. 3:17model. If we switch to video in the
  113. 3:20generation model, we can then click on
  114. 3:22the video models and we can scroll down
  115. 3:25and we can select Creatify Aurora.
  116. 3:27Within Creatify Aurora, we have a bunch
  117. 3:29of default avatars that we can choose.
  118. 3:32And so let's say we decide to choose
  119. 3:33this one right here. I could then click
  120. 3:35on speech and here we can go and add the
  121. 3:39audio that we've just generated. So if I
  122. 3:41click the plus I can use this as the
  123. 3:42speech and now I can simply click
  124. 3:44generate and Creatify Aurora is then
  125. 3:47going to create an AI avatar video of
  126. 3:49the default avatar plus the audio that
  127. 3:51I've just given it which is my voice
  128. 3:53clone. But if you can't find an AI
  129. 3:54avatar you like, you could then actually
  130. 3:56go and create your own simply by
  131. 3:58prompting the avatar that you want in
  132. 4:00[music] the image generation model with
  133. 4:03any image generation model. And so we
  134. 4:05could select [music] a model like Nano
  135. 4:07Banana Pro here. I'm going to go ahead
  136. 4:09and then use this prompt. And this
  137. 4:11prompt is just to give you a little bit
  138. 4:12of inspiration, but you could go ahead
  139. 4:14and describe any AI avatar for any
  140. 4:16[music] use case that you like. So, it
  141. 4:18could be for a construction site, it
  142. 4:20could be for a scientist, it could be
  143. 4:21for a doctor. And you simply type out
  144. 4:23what you want. Here, I want to change
  145. 4:25the resolution to at least 2K, but I
  146. 4:27would recommend even going up to 4K
  147. 4:29because the better the input of the
  148. 4:30image we give to AI video generation
  149. 4:33models, the [music] better the output.
  150. 4:34And then I'm going to choose four
  151. 4:36generations just so I've got a choice to
  152. 4:37choose from. And then I simply click
  153. 4:40generate. And so, as you can see, now
  154. 4:41we've got four different AI avatars that
  155. 4:43we could choose from. And if you didn't
  156. 4:45like any of them, you could go and click
  157. 4:46on the prompt again, regenerate to
  158. 4:48iterate on top of it. But if you did,
  159. 4:50what we could do is switch back to video
  160. 4:52and then within the Creatify Aurora
  161. 4:54model, we could simply drag and drop the
  162. 4:56one that we like into the AI avatar
  163. 4:58space. And we could scroll down to our
  164. 5:00previous generation and simply click on
  165. 5:02the audio again to use the same one and
  166. 5:04then click generate. But for this use
  167. 5:06case, I want to make an AI avatar that
  168. 5:08looks and [music] sounds 100% like me.
  169. 5:10We already have the voice. So now all we
  170. 5:12need is a photo and an image of me. Now
  171. 5:14if I was to click on avatar and then I
  172. 5:17click on upload, I could open up an
  173. 5:19image of me and then simply click
  174. 5:21generate. And now I have an AI avatar
  175. 5:24video that looks [music] and sounds like
  176. 5:25me. But the problem with this video is
  177. 5:27that the image is one by one. And I also
  178. 5:29don't like the layout on the backdrop.
  179. 5:30It doesn't feel like it was recorded and
  180. 5:33shot. It looks like we've turned an
  181. 5:34image into a video. And so what we could
  182. 5:37do if we go back to the image generation
  183. 5:39model, I can paste in this [music]
  184. 5:41prompt this time. And so as you can see
  185. 5:43at the start, I've got use my likeness
  186. 5:44from the uploaded reference image. And
  187. 5:46then I'm going to upload that same image
  188. 5:48as the reference image [music] with Nano
  189. 5:50Banana Pro. And then I'm going to change
  190. 5:52the aspect ratio to 4K because again, I
  191. 5:54want a really good input for the AI
  192. 5:56video models. And now I've got a short
  193. 5:59description of what I want the setting
  194. 6:01to be like. So I'm putting myself in a
  195. 6:02[music] new environment. And here I
  196. 6:04could change my clothes. I can change
  197. 6:06the background. I can change the
  198. 6:07lighting. [music] Anything I like. And
  199. 6:09by the way, let me know in the comments
  200. 6:10down below what type of AI avatar you
  201. 6:13are making. I'm very curious and I would
  202. 6:14love to see what you guys make. And
  203. 6:16before I click generate, I do want to
  204. 6:17mention that you can adjust the aspect
  205. 6:19ratio. And so here I want to turn it
  206. 6:20into a 16x9 video, but we could also
  207. 6:22make this [music] a 9x6 video. If you
  208. 6:25were making a vertical short form
  209. 6:28Instagram reel or Tik Tok, this would be
  210. 6:31perfect. And if you want something a
  211. 6:32little bit more UGC looking, we actually
  212. 6:34have a full tutorial on how to make UGC
  213. 6:37videos with AI, and we'll leave that in
  214. 6:39the description down below as well. And
  215. 6:41so now that I've got my reference image
  216. 6:42and the prompt, we simply [music] click
  217. 6:44generate. And so, as you can see, now
  218. 6:46I've got four still images that look
  219. 6:48100% like me that I can use as my AI
  220. 6:51avatar. And so, I really like this one.
  221. 6:52The environment looks so much better
  222. 6:54than the previous one, and it looks like
  223. 6:56a YouTube video, which is the aesthetic
  224. 6:59I was going for. [music] And so then
  225. 7:00again, I could go back to Creatify
  226. 7:02Aurora and I could drag and drop this
  227. 7:05one in [music] as the AI avatar image
  228. 7:08that I want to use. And then again go
  229. 7:10and add the speech simply by clicking on
  230. 7:13the previous audio that I used. And I
  231. 7:16would then go ahead and [music] click
  232. 7:18generate. And now I've got a custom AI
  233. 7:20avatar that still looks 100% [music]
  234. 7:22like me and still sounds 100% like me,
  235. 7:25except I'm in an environment that I was
  236. 7:27never [music] in. And right now, I've
  237. 7:29been using Creatify Aurora for all of my
  238. 7:31generations. But I do want to talk about
  239. 7:32a couple of other models and their
  240. 7:34limitations to help you pick which one
  241. 7:36you can use. Because if you click here,
  242. 7:38as you can see, there's a bunch of
  243. 7:40different AI models that you can choose.
  244. 7:42And we've been using some from the
  245. 7:43lipsync category, [music] but there's
  246. 7:45three that I want to talk about. The one
  247. 7:46that I've been using for all the
  248. 7:47generations so far has been created by
  249. 7:49Aurora. And the reason I like this model
  250. 7:51is that it has [music] a maximum length
  251. 7:52of 60 seconds. However, the resolution
  252. 7:55is capped to a maximum of 720p, but it's
  253. 7:58very expressive and it's pretty good for
  254. 8:00stable talking avatars. [music] The
  255. 8:02minus of it is just that you always need
  256. 8:03to upscale it afterwards, which you can
  257. 8:05also do inside of 11 Labs as well. The
  258. 8:07next one is Omnihuman 1.5. This one's
  259. 8:10nice. It's got a maximum resolution of
  260. 8:121080p, but the duration is capped at 30
  261. 8:15seconds. It is really good at [music]
  262. 8:17realistic faces though, but the clips
  263. 8:20are a little bit shorter and there's a
  264. 8:21little less movement and expressiveness
  265. 8:23in the face. And then finally, we've got
  266. 8:26LTX audio to video, which recently came
  267. 8:28out. [music] And this one, you can do
  268. 8:30resolutions of 720p and 1080p, but the
  269. 8:33maximum length is 20 seconds. However,
  270. 8:36this is an audiodriven AI [music] video
  271. 8:38model, and so it's not strictly
  272. 8:40lip-sync. This means that you can add a
  273. 8:42custom prompt with it to guide what's
  274. 8:44happening inside [music] of the video.
  275. 8:46and it will take into account your
  276. 8:47prompt and the audio as [music] well as
  277. 8:49the AI avatar start frame that you've
  278. 8:52given it and you can get some really
  279. 8:53cool results with this. And if you think
  280. 8:55about it, there are so many different
  281. 8:56ways that you can use this AI avatar
  282. 8:58content creation workflow in your
  283. 8:59creative processes. And let me show you
  284. 9:01a couple of cool things. First of all,
  285. 9:02if you didn't want to create a voice
  286. 9:04clone, you could actually upload your
  287. 9:06own voice recording. And so you're
  288. 9:08essentially lip-syncing a real audio
  289. 9:10recording of your voice to an AI avatar.
  290. 9:13And that's one workflow that you could
  291. 9:14do. But then if you think about it, you
  292. 9:16could take it even further and here you
  293. 9:18could go to voice changer. You could
  294. 9:20then record your own audio directly into
  295. 9:22voice changer and [music] you could then
  296. 9:24swap it to a new voice. So here right
  297. 9:27now I'm recording. And then if I click
  298. 9:29stop, I could then go to the voices
  299. 9:31section and change it to any voice I
  300. 9:34like and get a completely new sounding
  301. 9:36version of the recording I've just done
  302. 9:38and lip-s sync that with an invented AI
  303. 9:41character to get some really cool
  304. 9:43results. That is just a few of the many
  305. 9:46ways you can use the 11 Labs creative
  306. 9:49platform.
  307. 9:50>> And now you know how to create AI
  308. 9:52avatars with 11 Labs that look and sound
  309. 9:54100% like you. If you have any
  310. 9:56questions, let us know in the comment
  311. 9:57section down below. And if you enjoyed
  312. 9:59this video, please hit that like button
  313. 10:00and don't forget to subscribe. Thanks
  314. 10:02for watching.

About this transcript

This page contains the full transcript of Create Custom Realistic AI Avatars That Look & Sound 100% Like You (Full Workflow) by ElevenLabs, generated from the public captions YouTube serves with the video. The transcript has 2,333 words across 314 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.