I Automated YouTube Shorts with Python

[00:00] when I was scrolling through YouTube

[00:01] shorts yesterday I came across these

[00:03] videos seeing random facts about the

[00:05] world here's an example meet the

[00:06] immortal jellyfish a mesmerizing

[00:08] creature that defies the boundaries of

[00:10] life and death and I just thought how

[00:12] easy these would be to automate with

[00:13] literally little to no effort here's how

[00:16] that went I wanted to use Python to do

[00:18] this project since you can easily edit

[00:19] videos with it using a library called

[00:22] movie pie and because this is

[00:23] essentially free money method I named it

[00:25] money printer first of all I created a

[00:27] very simple front end where you can

[00:29] enter your video idea and it will

[00:30] automatically generate the video for you

[00:33] since I already know how to use the Bing

[00:35] version of cgt with python from previous

[00:37] projects I let it create a video script

[00:39] based on the provided user idea now that

[00:42] I have the script next up is to find

[00:43] good stock videos based on Search terms

[00:46] that jpt also generated for me for this

[00:48] I used the pixels API which is very easy

[00:50] to use and also free there is one

[00:52] problem however I don't know the

[00:54] resolution of the videos that the pixels

[00:56] API provides me with so I need to

[00:57] manually resize them using movie pie

[00:59] which is a pain since most of the time

[01:01] it simply renders this weird TV glitch

[01:03] screen but somehow I managed to fix it

[01:05] by cropping the video clips instead of

[01:07] resizing them now I had to find a form

[01:09] of Texas speech which didn't sound too

[01:11] robotic but I obviously don't have

[01:13] enough money for the 11 Labs API so we

[01:15] went with a Tik Tok TTS unofficial API

[01:18] from GitHub and it was converting the

[01:19] sentences in my script just fine there's

[01:21] a platform called assembly Ai and this

[01:23] is what I used to get a transcript of my

[01:25] just generated Texas speech as it would

[01:27] include timestamps and it can save the

[01:29] sub it in the SRT format movie piie

[01:32] comes in clutch Again by letting me burn

[01:34] subtitles into the video easily by

[01:36] concatenating the original video file

[01:38] with a subti clip class instance which

[01:40] just needs the path to my just exported

[01:43] SRT file with that the application is

[01:45] basically finished Let's test it out

[01:47] okay let's run our back end all right

[01:49] should be running now let's click on

[01:52] generate and now wait right now it's

[01:55] generating or generating our

[01:58] script it's generated it now these are

[02:01] the Search terms which we are going to

[02:03] use with the pixels API here are the

[02:05] videos it shows now it's downloading

[02:08] those okay it prints the script again

[02:12] now it's completing the text of speech

[02:14] using the Tik Tok API or unofficial API

[02:18] all right now it's generated the

[02:19] subtitles with the assembly Ai and last

[02:23] but not least it's combining everything

[02:24] together and once this is done we should

[02:27] have our output. MP4 I'm going to speed

[02:30] up this process a little bit cuz it's

[02:31] boring all right it seems like we still

[02:34] have to do something so I'm going to

[02:36] take that sacrifice because I'd rather

[02:38] wait a little while than edit all these

[02:40] videos myself so I'm fine with that

[02:43] great now it's done with that said let's

[02:45] have a look at our video the Amazon

[02:48] rainforest is the largest tropical

[02:50] rainforest in the world covering an area

[02:52] of approximately 5.5 million square km

[02:56] it is home to a diverse range of plant

[02:58] and animal species many of which are

[03:01] found nowhere else on Earth this looks

[03:03] good enough just so you don't think this

[03:05] was just some useless project I actually

[03:07] tested these clips on a channel and here

[03:09] are the

[03:11] results as you can see I did get a

[03:14] couple 1K views shorts which is pretty

[03:16] impressive for an automation bot in my

[03:18] opinion if you guys want the code like

[03:20] And subscribe as I'll be posting it on

[03:22] GitHub in the upcoming days till next

[03:28] time oh