1
00:00:00,000 --> 00:00:01,000
Hello guys.

2
00:00:01,000 --> 00:00:07,000
Welcome to this crash course of creating multiple AI agents for real world use cases using q AI if you

3
00:00:07,000 --> 00:00:08,000
haven't heard about this platform.

4
00:00:08,000 --> 00:00:14,000
So Q AI is an agent framework, uh, which will actually help you to create multiple agents for various

5
00:00:14,000 --> 00:00:16,000
amazing use cases.

6
00:00:16,000 --> 00:00:20,000
Um, so in this video, I'll be talking about this entire platform.

7
00:00:20,000 --> 00:00:24,000
Along with that, I will also show you with an amazing use cases.

8
00:00:24,000 --> 00:00:29,000
Uh, I will try to create a complete end to end use case and how I can use this Q AI platform or framework

9
00:00:30,000 --> 00:00:33,000
to develop multiple AI agents and where it will basically be required.

10
00:00:33,000 --> 00:00:34,000
See in long chain.

11
00:00:34,000 --> 00:00:37,000
Also you have an option to create agents, right?

12
00:00:37,000 --> 00:00:43,000
But here the main thing with respect to AI is that here your agent will be able to communicate with

13
00:00:43,000 --> 00:00:48,000
each other, right, in an efficient way so that we will be able to make or will be able to implement

14
00:00:48,000 --> 00:00:50,000
the task much more in an efficient way.

15
00:00:50,000 --> 00:00:54,000
Okay, so this is the entire web page of the crew.

16
00:00:54,000 --> 00:00:54,000
I.

17
00:00:54,000 --> 00:00:57,000
And right now many, many people are using it.

18
00:00:57,000 --> 00:00:59,000
They are creating multiple agent crews.

19
00:00:59,000 --> 00:01:02,000
You can see over here in last seven days all these things are there.

20
00:01:02,000 --> 00:01:05,000
You'll also be able to work with the open source tools that is already provided over here.

21
00:01:05,000 --> 00:01:10,000
And in this video I'll be also talking about that and showing you with the most practical use case.

22
00:01:10,000 --> 00:01:13,000
Um, so let me just go ahead and let me just start.

23
00:01:13,000 --> 00:01:16,000
First of all, we'll understand what kind of use case I will be solving.

24
00:01:16,000 --> 00:01:16,000
Okay.

25
00:01:16,000 --> 00:01:20,000
So let's say I have a project over here and everybody knows that I have a YouTube channel.

26
00:01:20,000 --> 00:01:24,000
Now in my YouTube channel, there are more than 1900 plus videos.

27
00:01:24,000 --> 00:01:29,000
Um, now, what I really want is that, uh, for every video, I need to have a blog page.

28
00:01:29,000 --> 00:01:32,000
Let's say I want to probably create a blog platform right now.

29
00:01:32,000 --> 00:01:34,000
I want to create a blog platform.

30
00:01:34,000 --> 00:01:35,000
Just let's imagine.

31
00:01:35,000 --> 00:01:38,000
So what I will do, I will take up every video of mine, probably based on the content.

32
00:01:38,000 --> 00:01:43,000
I will go ahead and write my entire blog page itself, Write whatever content is required in that blog

33
00:01:43,000 --> 00:01:43,000
page.

34
00:01:43,000 --> 00:01:46,000
Now, this task is a really tedious task.

35
00:01:46,000 --> 00:01:51,000
Right now, what I'm actually going to do is that I'm going to automate this entire thing with the help

36
00:01:51,000 --> 00:01:56,000
of AI framework, where this blog platform will be automatically created from my all my YouTube videos.

37
00:01:56,000 --> 00:02:01,000
With respect to all the things that I have said in my YouTube videos, that is most important.

38
00:02:01,000 --> 00:02:05,000
Okay, so with the help of Q AI, what we are really going to do is that first of all, let's say that

39
00:02:05,000 --> 00:02:12,000
if a user queries any kind of, uh, videos with respect to I, let's say I'm querying about Q I over

40
00:02:12,000 --> 00:02:13,000
here.

41
00:02:13,000 --> 00:02:18,000
So what it should happen is that it should go to my YouTube channel, pick that particular video, extract

42
00:02:18,000 --> 00:02:22,000
that entire content and based on that particular content, summarize that information, put it into

43
00:02:22,000 --> 00:02:23,000
blog page.

44
00:02:23,000 --> 00:02:23,000
Right.

45
00:02:24,000 --> 00:02:30,000
Now, obviously, I know that I have created more than 1900 plus videos now if I really want to probably

46
00:02:30,000 --> 00:02:32,000
work with multiple people over here.

47
00:02:32,000 --> 00:02:33,000
So I need to have a content writer.

48
00:02:33,000 --> 00:02:39,000
I need to have a researcher, right, who will probably go and explore every video in my YouTube channel

49
00:02:39,000 --> 00:02:43,000
based on the query it is writing, and then it will go and validate that content.

50
00:02:43,000 --> 00:02:47,000
And after probably validating the content, I need to have a content writer separately who will specifically

51
00:02:47,000 --> 00:02:49,000
write this particular blog page.

52
00:02:49,000 --> 00:02:54,000
Also right now here, um, just to understand, for just one video it is fine.

53
00:02:54,000 --> 00:02:59,000
But if I have 1900 plus videos, it is going to take a lot of time because validating is also there.

54
00:02:59,000 --> 00:03:03,000
I need to probably go and check the content proofing, do multiple things over there, right?

55
00:03:03,000 --> 00:03:10,000
So with the crew AI, we can automate this thing completely in a whisker of time.

56
00:03:10,000 --> 00:03:14,000
Uh, and that that is where we will understand how this AI agents.

57
00:03:14,000 --> 00:03:14,000
Okay.

58
00:03:14,000 --> 00:03:18,000
And here also you can see, right, if I'm involving multiple people, they also need to communicate

59
00:03:18,000 --> 00:03:19,000
with each other.

60
00:03:19,000 --> 00:03:22,000
Then again have there may be some kind of communication gap.

61
00:03:22,000 --> 00:03:24,000
You know how they are specifically going to work.

62
00:03:24,000 --> 00:03:29,000
So with respect to that you are going to take a lot of time over here right now with the help of creo

63
00:03:29,000 --> 00:03:34,000
I um so creo I if I talk about they are three main important components.

64
00:03:34,000 --> 00:03:37,000
One is the agents, one is the task and one is the tools.

65
00:03:37,000 --> 00:03:37,000
Okay.

66
00:03:37,000 --> 00:03:42,000
Now here if I probably consider this specific use cases as I'm saying.

67
00:03:42,000 --> 00:03:44,000
first of all, we need to explore my YouTube channel, right?

68
00:03:44,000 --> 00:03:49,000
So I need to have a researcher over here who will probably go and query any query in my YouTube video,

69
00:03:49,000 --> 00:03:52,000
and it will probably bring that particular YouTube video.

70
00:03:52,000 --> 00:03:52,000
Okay.

71
00:03:52,000 --> 00:03:55,000
After seeing that particular YouTube video, he needs to listen.

72
00:03:55,000 --> 00:04:00,000
He needs to probably hear out what what you have about that entire YouTube video itself.

73
00:04:00,000 --> 00:04:05,000
So here also we require an expert right expert who is good at data science, who is good at data analysts

74
00:04:05,000 --> 00:04:09,000
because my entire channel is based on data science itself, right?

75
00:04:09,000 --> 00:04:14,000
I so here obviously I need an expert person who will probably go ahead and watch each and every video

76
00:04:14,000 --> 00:04:16,000
and retrieve the content out of it.

77
00:04:16,000 --> 00:04:17,000
Right.

78
00:04:17,000 --> 00:04:20,000
Then after that, I need to pass it to my content writer.

79
00:04:20,000 --> 00:04:20,000
Right.

80
00:04:20,000 --> 00:04:22,000
So content writer.

81
00:04:22,000 --> 00:04:25,000
So here I have an domain expert person, right?

82
00:04:25,000 --> 00:04:26,000
It can be a data scientist.

83
00:04:26,000 --> 00:04:27,000
It can be an analyst.

84
00:04:27,000 --> 00:04:28,000
Right.

85
00:04:28,000 --> 00:04:34,000
So this two roles here you can basically consider in queue I these are nothing but these are agents

86
00:04:34,000 --> 00:04:35,000
okay.

87
00:04:35,000 --> 00:04:38,000
So we are specifically going to take consider this as agents.

88
00:04:38,000 --> 00:04:44,000
So agents are none other than the people who are having some kind of experience with respect to the

89
00:04:44,000 --> 00:04:44,000
work.

90
00:04:44,000 --> 00:04:45,000
Right.

91
00:04:45,000 --> 00:04:47,000
It can be a domain expertise like in data science domain.

92
00:04:47,000 --> 00:04:49,000
It can be data scientist, data analyst.

93
00:04:49,000 --> 00:04:51,000
It can be a content writer.

94
00:04:51,000 --> 00:04:51,000
Right.

95
00:04:51,000 --> 00:04:52,000
So it can be anyone.

96
00:04:52,000 --> 00:04:52,000
Right.

97
00:04:52,000 --> 00:04:54,000
So this basically becomes an agent.

98
00:04:54,000 --> 00:04:58,000
So in the Q I the first component is specifically agents.

99
00:04:58,000 --> 00:05:02,000
The second component that we specifically use in Q is task.

100
00:05:02,000 --> 00:05:06,000
Right now each and every agent does some kind of task.

101
00:05:06,000 --> 00:05:11,000
Let's say the domain expert people will probably go ahead and watch this particular YouTube video,

102
00:05:11,000 --> 00:05:15,000
try to translate, uh, take out the entire information that I've told in the YouTube video itself.

103
00:05:15,000 --> 00:05:15,000
Right.

104
00:05:15,000 --> 00:05:20,000
So, uh, that basically becomes the task for this particular agent, right?

105
00:05:20,000 --> 00:05:22,000
Similarly, content writer is the agent.

106
00:05:22,000 --> 00:05:24,000
Its task is to write the content.

107
00:05:24,000 --> 00:05:25,000
Right.

108
00:05:25,000 --> 00:05:26,000
So that can basically be a task.

109
00:05:26,000 --> 00:05:29,000
So this is the most important second component.

110
00:05:29,000 --> 00:05:33,000
Like what task a agent is specifically doing that also we need to define.

111
00:05:33,000 --> 00:05:37,000
And the third thing over here is about tools okay.

112
00:05:37,000 --> 00:05:40,000
Let's say uh the domain expertise is over here.

113
00:05:40,000 --> 00:05:41,000
I do not have a domain expertise.

114
00:05:41,000 --> 00:05:42,000
Let's consider.

115
00:05:42,000 --> 00:05:47,000
So if that domain expert is one some kind of help, right.

116
00:05:47,000 --> 00:05:52,000
Uh, like, uh, let's say once it is searching from the YouTube, uh, video itself.

117
00:05:52,000 --> 00:05:52,000
Right.

118
00:05:52,000 --> 00:05:57,000
Uh, any of my video, it is probably searching, let's say that it wants to get the transcript of the

119
00:05:57,000 --> 00:05:57,000
video.

120
00:05:57,000 --> 00:06:02,000
So what may what may happen is that this person may use some kind of tool to do that.

121
00:06:02,000 --> 00:06:03,000
It can be a third party tool.

122
00:06:03,000 --> 00:06:04,000
It can be an API.

123
00:06:04,000 --> 00:06:05,000
It can be anything as such.

124
00:06:05,000 --> 00:06:09,000
In this particular case, uh, let's say I want to get the transcript of the my YouTube video.

125
00:06:09,000 --> 00:06:11,000
Then I may use a transcriber.

126
00:06:11,000 --> 00:06:11,000
Right.

127
00:06:11,000 --> 00:06:14,000
And that transcriber can be a third party tool itself.

128
00:06:14,000 --> 00:06:14,000
Right.

129
00:06:14,000 --> 00:06:17,000
And that is where a tool comes into existence.

130
00:06:17,000 --> 00:06:24,000
Like with what will be the main, uh, what will be the main way of probably let's say that if I have

131
00:06:24,000 --> 00:06:30,000
some kind of dependency on some third party tools itself to explore this particular task, right.

132
00:06:30,000 --> 00:06:33,000
How can I perform this particular task with the help of this particular tool?

133
00:06:33,000 --> 00:06:39,000
So I may also have a separate tool which will be able to, uh, provide me the transcription of the

134
00:06:39,000 --> 00:06:40,000
entire video.

135
00:06:40,000 --> 00:06:41,000
Right.

136
00:06:41,000 --> 00:06:43,000
So that is where we can specifically use the tool.

137
00:06:43,000 --> 00:06:47,000
Other tool can be okay, I want to probably do a Google search API right.

138
00:06:47,000 --> 00:06:47,000
Google search.

139
00:06:47,000 --> 00:06:49,000
So this can be another tool.

140
00:06:49,000 --> 00:06:49,000
Right.

141
00:06:49,000 --> 00:06:53,000
So similarly multiple tools will basically be there which we can actually use.

142
00:06:53,000 --> 00:06:56,000
So agent will have some specific task.

143
00:06:56,000 --> 00:06:58,000
And this task can be performed by a specific tool.

144
00:06:58,000 --> 00:06:59,000
Okay.

145
00:06:59,000 --> 00:07:01,000
So this is how it actually works.

146
00:07:01,000 --> 00:07:04,000
Now similarly here you can also see I have created two agents.

147
00:07:04,000 --> 00:07:07,000
One is researcher, one is content writer because that is what I actually require.

148
00:07:07,000 --> 00:07:11,000
Researcher will explore the videos from this particular videos.

149
00:07:11,000 --> 00:07:15,000
It may use some kind of I.T tools to transcribe the entire content.

150
00:07:15,000 --> 00:07:15,000
Right.

151
00:07:15,000 --> 00:07:17,000
Analyze the entire content.

152
00:07:17,000 --> 00:07:21,000
Once that thing is done, then that particular researcher will pass it to the content writer because

153
00:07:21,000 --> 00:07:23,000
it now has the entire information.

154
00:07:23,000 --> 00:07:27,000
This entire info will be basically there after performing this particular task.

155
00:07:28,000 --> 00:07:32,000
Then once it provides it to the content writer, the content writer, what will happen based on the

156
00:07:32,000 --> 00:07:32,000
research?

157
00:07:32,000 --> 00:07:34,000
It will write the blog page, right?

158
00:07:34,000 --> 00:07:37,000
And finally, this will basically be my output, right?

159
00:07:37,000 --> 00:07:42,000
So here you can also see that interaction is there between the researcher and the content writer.

160
00:07:42,000 --> 00:07:43,000
Right.

161
00:07:43,000 --> 00:07:45,000
And this this process.

162
00:07:45,000 --> 00:07:46,000
Now see, this is one way.

163
00:07:46,000 --> 00:07:49,000
So this process is entirely called a sequential process.

164
00:07:49,000 --> 00:07:52,000
There are also other processes which is called as hierarchical processes.

165
00:07:52,000 --> 00:07:57,000
So here in the sequential process once the researcher completes his work, it is going to give that

166
00:07:57,000 --> 00:08:02,000
entire output to the content writer and the content writer further, based on the task it is assigned,

167
00:08:02,000 --> 00:08:04,000
it will probably go ahead and create the blog page.

168
00:08:04,000 --> 00:08:09,000
So I hope you got an idea about what are the main important components of creo I.

169
00:08:09,000 --> 00:08:12,000
One is the agent, one is the task and one is the tools.

170
00:08:12,000 --> 00:08:15,000
And based on this we can automate this entire use cases.

171
00:08:15,000 --> 00:08:18,000
Now let's go ahead and implement it practically.

172
00:08:18,000 --> 00:08:21,000
As I said in the agents I'm going to probably create two agents.

173
00:08:21,000 --> 00:08:23,000
One is researcher, one is the content writer.

174
00:08:23,000 --> 00:08:27,000
Then I'm going to go ahead and define the task for this specific agent.

175
00:08:27,000 --> 00:08:28,000
Like explore videos.

176
00:08:28,000 --> 00:08:30,000
It can be explore videos.

177
00:08:30,000 --> 00:08:31,000
It can be exploring.

178
00:08:31,000 --> 00:08:35,000
Another thing it can be probably exploring Google Search API, anything it can be.

179
00:08:35,000 --> 00:08:39,000
Now to complete this particular task, we have dependency on some tool which is called as y t tools

180
00:08:39,000 --> 00:08:45,000
because, uh, at the end of the day, I require the transcription or uh, of my entire YouTube video.

181
00:08:45,000 --> 00:08:47,000
So for that I may require a tool.

182
00:08:47,000 --> 00:08:50,000
And this can also be a custom tool, which you can also create by yourself.

183
00:08:50,000 --> 00:08:50,000
Right?

184
00:08:50,000 --> 00:08:55,000
So once we get this particular tool, we will be able to do this particular task.

185
00:08:55,000 --> 00:08:58,000
And after that we will be able to complete the researcher work.

186
00:08:58,000 --> 00:09:02,000
Once this researcher work is completed, we will pass this entire work to the content writer, because

187
00:09:02,000 --> 00:09:07,000
the content writer needs to write the entire blog page based on the research, right?

188
00:09:07,000 --> 00:09:12,000
So this entire process is also called as sequential process because once this is getting completed,

189
00:09:12,000 --> 00:09:14,000
the next step is to get, uh, to complete this.

190
00:09:14,000 --> 00:09:19,000
There are also other processing, uh, other processes like hierarchical process where parallelly also

191
00:09:19,000 --> 00:09:21,000
you can actually do this particular task.

192
00:09:21,000 --> 00:09:21,000
Okay.

193
00:09:21,000 --> 00:09:25,000
Now let's go ahead and implement this entire project completely from scratch.

194
00:09:25,000 --> 00:09:29,000
So guys, I have opened my, uh, new project over here in my VS code.

195
00:09:29,000 --> 00:09:30,000
I will go to the terminal.

196
00:09:30,000 --> 00:09:34,000
The first step is probably to create our conda environment.

197
00:09:34,000 --> 00:09:34,000
Okay.

198
00:09:34,000 --> 00:09:37,000
And this you really need to do it for every project.

199
00:09:37,000 --> 00:09:47,000
So I will go ahead and write conda create conda create minus p v n v python double equal to 3.10.

200
00:09:47,000 --> 00:09:49,000
So okay I will be taking 3.10.

201
00:09:49,000 --> 00:09:50,000
That is 3.10.

202
00:09:50,000 --> 00:09:56,000
Now after doing this uh let's say installation probably happen uh, once the installation actually happens

203
00:09:56,000 --> 00:09:58,000
on a new environment is basically created.

204
00:09:58,000 --> 00:10:03,000
What we are going to basically do is that we are going to create our requirement dot txt file.

205
00:10:03,000 --> 00:10:07,000
So let me just go ahead and write my requirement dot txt file.

206
00:10:07,000 --> 00:10:08,000
Okay.

207
00:10:08,000 --> 00:10:14,000
Uh now inside this requirement dot txt file I will be using some of libraries that I really need to

208
00:10:14,000 --> 00:10:14,000
install.

209
00:10:14,000 --> 00:10:15,000
One is the CRI.

210
00:10:15,000 --> 00:10:18,000
Then one I'm going to write Python.

211
00:10:18,000 --> 00:10:19,000
Okay.

212
00:10:19,000 --> 00:10:20,000
Right now I don't require this.

213
00:10:20,000 --> 00:10:26,000
So let me just go ahead and write creo underscore I sorry creo I underscore tools right.

214
00:10:26,000 --> 00:10:29,000
So these are the two important libraries I will be requiring.

215
00:10:29,000 --> 00:10:31,000
So let me save it.

216
00:10:31,000 --> 00:10:35,000
Now this environment is basically getting created or it has got created.

217
00:10:35,000 --> 00:10:38,000
So I will go ahead and activate v and v okay.

218
00:10:38,000 --> 00:10:40,000
So conda activate v and v.

219
00:10:40,000 --> 00:10:44,000
Now let me quickly go ahead and let me do one thing guys.

220
00:10:44,000 --> 00:10:47,000
Let me hide my face so that you will be able to clearly see this okay.

221
00:10:48,000 --> 00:10:53,000
Now let me just quickly go ahead and write pip install minus r requirement dot txt.

222
00:10:53,000 --> 00:10:57,000
And this all both the requirements will get installed okay.

223
00:10:57,000 --> 00:11:03,000
Okay, so once this installation is basically happening uh, we will continue our task.

224
00:11:03,000 --> 00:11:06,000
The first of all, as you know we need to create our agents.

225
00:11:06,000 --> 00:11:09,000
So agents dot p y I will go ahead and create it.

226
00:11:09,000 --> 00:11:12,000
The next will basically be tools dot p y.

227
00:11:12,000 --> 00:11:13,000
Okay.

228
00:11:14,000 --> 00:11:18,000
Uh, agent tools and I also require task okay.

229
00:11:18,000 --> 00:11:20,000
Task dot p y.

230
00:11:20,000 --> 00:11:23,000
So these three components are we really need to create.

231
00:11:23,000 --> 00:11:27,000
So first of all I will go ahead and create my agents.

232
00:11:27,000 --> 00:11:28,000
Now for creating the agents.

233
00:11:28,000 --> 00:11:31,000
First of all I will go ahead and import from Korea.

234
00:11:31,000 --> 00:11:35,000
I import import agent.

235
00:11:36,000 --> 00:11:37,000
Okay.

236
00:11:37,000 --> 00:11:44,000
Now uh, since, uh, you know that, uh, we really need to create some kind of agent over here, right?

237
00:11:44,000 --> 00:11:46,000
And for creating an agent.

238
00:11:46,000 --> 00:11:48,000
So first of all, what all things I will create.

239
00:11:48,000 --> 00:11:57,000
So I will probably create a, uh, senior blog content researcher.

240
00:11:57,000 --> 00:11:58,000
Okay.

241
00:11:58,000 --> 00:12:03,000
So this will basically be my first researcher, which I will probably create, which will be an agent

242
00:12:03,000 --> 00:12:04,000
who will be doing my task.

243
00:12:04,000 --> 00:12:05,000
Okay.

244
00:12:05,000 --> 00:12:08,000
So these are like people who will be handling all my tasks.

245
00:12:08,000 --> 00:12:08,000
Okay.

246
00:12:08,000 --> 00:12:11,000
Now let me quickly go ahead and create my researcher.

247
00:12:11,000 --> 00:12:14,000
So this researcher will be my blog researcher.

248
00:12:14,000 --> 00:12:15,000
Okay.

249
00:12:15,000 --> 00:12:18,000
And this blog researcher will be of type agent.

250
00:12:18,000 --> 00:12:19,000
Okay.

251
00:12:19,000 --> 00:12:24,000
Now with respect to agent, uh, there are some important parameters that we really need to use.

252
00:12:24,000 --> 00:12:25,000
Okay.

253
00:12:25,000 --> 00:12:28,000
First parameter we really need to give is role.

254
00:12:28,000 --> 00:12:33,000
What kind of role it is basically doing or what kind of role it needs to do it okay.

255
00:12:33,000 --> 00:12:37,000
So here I'm going to basically say that okay, these are nothing, but they have to probably be a block

256
00:12:37,000 --> 00:12:45,000
creator or block researcher from YouTube videos okay, YouTube videos.

257
00:12:46,000 --> 00:12:49,000
So these are some default information that we really need to give.

258
00:12:49,000 --> 00:12:50,000
One is the role.

259
00:12:50,000 --> 00:12:51,000
The other one is goal.

260
00:12:51,000 --> 00:12:51,000
Okay.

261
00:12:51,000 --> 00:12:55,000
So here we also need to specify the uh goal over here.

262
00:12:56,000 --> 00:13:06,000
Uh here we can basically write get the relevant, relevant video, uh, get the relevant video, um,

263
00:13:07,000 --> 00:13:14,000
content for the topic, whatever topic I say.

264
00:13:15,000 --> 00:13:16,000
Okay.

265
00:13:16,000 --> 00:13:18,000
From white channel.

266
00:13:18,000 --> 00:13:18,000
Okay.

267
00:13:18,000 --> 00:13:22,000
So this will basically be my role for this particular agent.

268
00:13:22,000 --> 00:13:26,000
You need to get the relevant video, uh, content for the topic.

269
00:13:26,000 --> 00:13:28,000
This from my, uh, this one.

270
00:13:28,000 --> 00:13:28,000
Okay.

271
00:13:28,000 --> 00:13:29,000
Name.

272
00:13:29,000 --> 00:13:32,000
I can basically, uh, right over here description so I can.

273
00:13:32,000 --> 00:13:32,000
Right.

274
00:13:32,000 --> 00:13:36,000
So let me just go ahead and uh, uh, write the name, but let it be.

275
00:13:36,000 --> 00:13:42,000
I don't require this too, but it is getting suggested by the, uh, Amazon whisper that I have specifically

276
00:13:42,000 --> 00:13:43,000
used over here, but it's okay.

277
00:13:43,000 --> 00:13:44,000
I don't want this.

278
00:13:44,000 --> 00:13:48,000
Then I will go ahead and set up my verbose.

279
00:13:48,000 --> 00:13:53,000
So verbose will be true, which will be able to see some information out over here.

280
00:13:53,000 --> 00:13:56,000
We're also going to set one parameter which is called as memory true.

281
00:13:56,000 --> 00:13:58,000
So which will be initialized with some memory.

282
00:13:58,000 --> 00:13:59,000
And we will go.

283
00:13:59,000 --> 00:14:03,000
Also go ahead and write some backstory about this particular agent okay.

284
00:14:03,000 --> 00:14:05,000
So let's go ahead and define some backstory.

285
00:14:05,000 --> 00:14:09,000
So uh I will keep a backstory something like this.

286
00:14:09,000 --> 00:14:12,000
This person or this agent is an expert.

287
00:14:13,000 --> 00:14:15,000
Okay, see?

288
00:14:15,000 --> 00:14:15,000
See this?

289
00:14:15,000 --> 00:14:17,000
Okay, so I will just go ahead and write.

290
00:14:17,000 --> 00:14:22,000
This person is expert in understanding videos in AI, data science, machine learning and gen.

291
00:14:22,000 --> 00:14:23,000
I am providing suggestions.

292
00:14:23,000 --> 00:14:24,000
Okay.

293
00:14:24,000 --> 00:14:27,000
So this will basically be my back story.

294
00:14:27,000 --> 00:14:32,000
And then the third thing that I actually require is my tool, whether I'm going to use some tools or

295
00:14:32,000 --> 00:14:33,000
not.

296
00:14:33,000 --> 00:14:36,000
So here I'm going to basically define my tool.

297
00:14:36,000 --> 00:14:39,000
And right now I'll keep the tool empty because I have not created any tool.

298
00:14:39,000 --> 00:14:40,000
Okay.

299
00:14:40,000 --> 00:14:43,000
And then further I will also say allow delegation.

300
00:14:43,000 --> 00:14:50,000
Delegation basically means will I be transferring after whatever work that I do or this agent does to

301
00:14:50,000 --> 00:14:51,000
someone else.

302
00:14:51,000 --> 00:14:54,000
So we will set this allow delegation to true.

303
00:14:54,000 --> 00:14:54,000
Okay.

304
00:14:54,000 --> 00:14:59,000
So this is all the default parameters that we really need to write for an agent.

305
00:14:59,000 --> 00:14:59,000
Okay.

306
00:14:59,000 --> 00:15:02,000
Uh, and based on this you can create anything.

307
00:15:02,000 --> 00:15:07,000
As such, this topic will be the topic that I will probably be giving for whichever topic I really want

308
00:15:07,000 --> 00:15:09,000
to create my blog from my YouTube videos.

309
00:15:09,000 --> 00:15:12,000
Okay, now the third thing.

310
00:15:12,000 --> 00:15:12,000
Third.

311
00:15:12,000 --> 00:15:12,000
Right.

312
00:15:12,000 --> 00:15:14,000
Let's go ahead and define this.

313
00:15:14,000 --> 00:15:19,000
The second agent that we will be creating.

314
00:15:19,000 --> 00:15:28,000
So here I'm going to write a writer agent blog writer agent creating a senior blog writer agent okay.

315
00:15:29,000 --> 00:15:32,000
With with white tool.

316
00:15:32,000 --> 00:15:33,000
Okay.

317
00:15:33,000 --> 00:15:38,000
So right now I'm also going to create my white tool, which I will also show you how to probably create.

318
00:15:38,000 --> 00:15:40,000
So this basically becomes my writer.

319
00:15:41,000 --> 00:15:44,000
Or I can also say it as blog underscore writer.

320
00:15:44,000 --> 00:15:48,000
And this blog writer will be again an agent type.

321
00:15:49,000 --> 00:15:50,000
Agent type.

322
00:15:50,000 --> 00:15:57,000
And here I'm just going to copy and paste, uh, some information with respect to roles, tools and

323
00:15:57,000 --> 00:15:57,000
all.

324
00:15:57,000 --> 00:15:58,000
Okay.

325
00:15:58,000 --> 00:16:00,000
So now let me copy it over here.

326
00:16:00,000 --> 00:16:06,000
And here you can probably see that I have written some of the information like role is nothing but blog

327
00:16:06,000 --> 00:16:06,000
writer.

328
00:16:06,000 --> 00:16:11,000
I can basically write it as something like this blog writer goal is narrate the compelling text stories

329
00:16:11,000 --> 00:16:20,000
about the video, or I'm basically going to write with respect to any topic over here from y t channel

330
00:16:21,000 --> 00:16:22,000
from y t channel.

331
00:16:22,000 --> 00:16:24,000
Okay, verbose will be true.

332
00:16:24,000 --> 00:16:25,000
Memory will be true.

333
00:16:25,000 --> 00:16:25,000
Back story.

334
00:16:25,000 --> 00:16:29,000
We have specifying some amazing back story with a fear of simplifying the complex topic.

335
00:16:29,000 --> 00:16:36,000
Your graph engaging narratives that captivate and educated, bringing new discoveries to light in accessible

336
00:16:36,000 --> 00:16:36,000
manner.

337
00:16:36,000 --> 00:16:41,000
Then tool I have not yet created and allow delegate because see, at the end of the day, a blog writer

338
00:16:41,000 --> 00:16:43,000
needs to just write the blog.

339
00:16:43,000 --> 00:16:46,000
So further, we do not have to delegate his work to someone else.

340
00:16:46,000 --> 00:16:48,000
So we are keeping this as false.

341
00:16:48,000 --> 00:16:50,000
Okay, so this is what is my agent.

342
00:16:50,000 --> 00:16:52,000
So two agents have actually created.

343
00:16:52,000 --> 00:16:55,000
One is my blog researcher and one is my blog writer.

344
00:16:55,000 --> 00:16:58,000
Now let's go ahead with the next step over here.

345
00:16:58,000 --> 00:17:01,000
Uh, and uh, let's go ahead and define my task.

346
00:17:02,000 --> 00:17:05,000
So task or tools, first of all we'll go ahead and define my tools.

347
00:17:05,000 --> 00:17:13,000
So for tools uh again if I go to my documentation page in career I so uh, get started.

348
00:17:13,000 --> 00:17:14,000
Let's see.

349
00:17:14,000 --> 00:17:15,000
Let's get started.

350
00:17:15,000 --> 00:17:18,000
So here if you go down there are multiple tools which it supports.

351
00:17:18,000 --> 00:17:19,000
Right.

352
00:17:19,000 --> 00:17:23,000
So here you can see Google search server search browser based web loader.

353
00:17:23,000 --> 00:17:25,000
So many different different tools are there.

354
00:17:25,000 --> 00:17:29,000
My, uh the tool that I have actually used is nothing but YouTube channel.

355
00:17:29,000 --> 00:17:30,000
Uh, search tool.

356
00:17:30,000 --> 00:17:31,000
Okay.

357
00:17:31,000 --> 00:17:33,000
So this will basically be searching my YouTube channel.

358
00:17:33,000 --> 00:17:34,000
Okay.

359
00:17:34,000 --> 00:17:36,000
So first of all it says pip install creo AI tools.

360
00:17:36,000 --> 00:17:38,000
I have already done that installation.

361
00:17:38,000 --> 00:17:41,000
If you remember I underscore tools okay.

362
00:17:41,000 --> 00:17:44,000
And this is how we explore any tool name okay.

363
00:17:44,000 --> 00:17:46,000
Sorry this YouTube search tool.

364
00:17:46,000 --> 00:17:50,000
So first of all we'll go ahead and import from Crewe Q I import YouTube search tool.

365
00:17:50,000 --> 00:17:53,000
So let me just go ahead and import this okay.

366
00:17:53,000 --> 00:18:01,000
And uh after that we will go ahead and initialize our YouTube channel, which YouTube channel we really

367
00:18:01,000 --> 00:18:02,000
want to explore.

368
00:18:02,000 --> 00:18:04,000
And I want my YouTube channel to explore.

369
00:18:04,000 --> 00:18:07,000
So I will just go ahead and write my YouTube channel name Krishna zero six.

370
00:18:07,000 --> 00:18:07,000
Okay.

371
00:18:07,000 --> 00:18:10,000
So this is uh, this is my tool.

372
00:18:10,000 --> 00:18:15,000
So this tool is responsible whenever I get any query, it should be searching from this particular YouTube

373
00:18:15,000 --> 00:18:15,000
channel.

374
00:18:15,000 --> 00:18:20,000
And this tool we will be using now since this tool is what we are going to use.

375
00:18:20,000 --> 00:18:22,000
So we need to update in our agents.

376
00:18:22,000 --> 00:18:27,000
So here I'm going to write from tool from tools Import okay.

377
00:18:27,000 --> 00:18:32,000
Instead of writing tool here I will go ahead and write my y t underscore two okay.

378
00:18:32,000 --> 00:18:37,000
So this variable I will go ahead and initialize in my agents dot p y Okay.

379
00:18:37,000 --> 00:18:40,000
And the same tool I will be using inside my tools.

380
00:18:41,000 --> 00:18:42,000
Now this is completed.

381
00:18:42,000 --> 00:18:43,000
Amazing right.

382
00:18:43,000 --> 00:18:46,000
So my agents and tools are actually completed.

383
00:18:46,000 --> 00:18:51,000
If you want any other tool like Serp API, Google Search API, you can just follow this particular page.

384
00:18:51,000 --> 00:18:51,000
Right.

385
00:18:51,000 --> 00:18:56,000
So uh, and don't worry as I will go ahead, I'll be creating more amazing videos.

386
00:18:56,000 --> 00:18:59,000
So let's say that I want to go ahead and do PDF rack search.

387
00:18:59,000 --> 00:19:03,000
So you can probably use this particular tool for doing that same process.

388
00:19:03,000 --> 00:19:05,000
Right to ask any question from that.

389
00:19:05,000 --> 00:19:10,000
But this is the most efficient way of creating any, um, you know, agents in an easy way.

390
00:19:10,000 --> 00:19:13,000
Now, the third thing that we really need to fill is our task.

391
00:19:13,000 --> 00:19:14,000
Okay.

392
00:19:14,000 --> 00:19:18,000
Now, since, uh, you know that we have created two agents, right?

393
00:19:18,000 --> 00:19:21,000
One is the blog researcher and one is the blog writer.

394
00:19:21,000 --> 00:19:23,000
So similarly, two task needs to be created over here.

395
00:19:23,000 --> 00:19:25,000
So I'm going to write from crew Crewe.

396
00:19:26,000 --> 00:19:33,000
I import task okay from tools.

397
00:19:35,000 --> 00:19:38,000
Here also we are going to use tools import tool.

398
00:19:38,000 --> 00:19:38,000
Okay.

399
00:19:38,000 --> 00:19:40,000
So these are the two things.

400
00:19:40,000 --> 00:19:50,000
And then uh I also need to probably call from agents import blog Researcher.

401
00:19:51,000 --> 00:19:52,000
Comma.

402
00:19:52,000 --> 00:19:53,000
Blog writer.

403
00:19:53,000 --> 00:19:54,000
So both the agents.

404
00:19:54,000 --> 00:19:56,000
Also I need to probably call.

405
00:19:56,000 --> 00:19:58,000
Okay, done.

406
00:19:59,000 --> 00:20:03,000
So let me just see the name agents, blog researcher and blog writer.

407
00:20:03,000 --> 00:20:07,000
So both the a both the agents have actually called over here.

408
00:20:07,000 --> 00:20:12,000
Now the first task will be nothing but the research task.

409
00:20:12,000 --> 00:20:12,000
Okay.

410
00:20:12,000 --> 00:20:16,000
Now in order to properly use this, I have to again use this task.

411
00:20:16,000 --> 00:20:17,000
And, uh, it is also very simple.

412
00:20:17,000 --> 00:20:21,000
So here I will probably go ahead and write and create my research task.

413
00:20:21,000 --> 00:20:25,000
So here you can see here some basic information is required inside this task.

414
00:20:25,000 --> 00:20:30,000
Class one is the description like identify the video okay.

415
00:20:30,000 --> 00:20:32,000
Get detailed information about the video from the channel.

416
00:20:32,000 --> 00:20:34,000
So these are the two information that I have given.

417
00:20:35,000 --> 00:20:39,000
Expected output is nothing but a comprehensive three paragraph long report based on the topic of the

418
00:20:39,000 --> 00:20:40,000
video content.

419
00:20:40,000 --> 00:20:44,000
And here I'm going to use this particular tool over here.

420
00:20:44,000 --> 00:20:48,000
But since this tool name is not there I will go ahead and use it tool.

421
00:20:48,000 --> 00:20:52,000
So let me copy paste it over here and let me copy paste it over here.

422
00:20:52,000 --> 00:20:55,000
So white underscore tool will be there.

423
00:20:56,000 --> 00:20:59,000
So here you can see that I have updated this researcher task.

424
00:20:59,000 --> 00:21:00,000
What tool I'm basically going to use.

425
00:21:00,000 --> 00:21:02,000
So white tool I'm going to use.

426
00:21:02,000 --> 00:21:05,000
and my agent will basically be my blog researcher.

427
00:21:05,000 --> 00:21:11,000
Okay, so these are the updated information that I have with respect to, uh, this entire research

428
00:21:11,000 --> 00:21:11,000
task.

429
00:21:11,000 --> 00:21:15,000
Similarly, I will go ahead and create my writing task.

430
00:21:15,000 --> 00:21:17,000
And again we will be using the writing class.

431
00:21:17,000 --> 00:21:20,000
So here you can see writing task task description.

432
00:21:20,000 --> 00:21:23,000
Get the info from the YouTube channel or topic this.

433
00:21:23,000 --> 00:21:26,000
Summarize the info about the video on this particular topic.

434
00:21:26,000 --> 00:21:26,000
Tools.

435
00:21:26,000 --> 00:21:29,000
I'm again going to use white tool here.

436
00:21:29,000 --> 00:21:31,000
Also we are going to specifically use white tool.

437
00:21:31,000 --> 00:21:36,000
Along with that, my agent will be nothing but my blog writer which will be doing the task.

438
00:21:37,000 --> 00:21:38,000
Async execution.

439
00:21:38,000 --> 00:21:40,000
Okay, there is one parameter which is called as async execution.

440
00:21:40,000 --> 00:21:42,000
Async async execution.

441
00:21:42,000 --> 00:21:46,000
If I set it to true then both this agent will be parallelly working.

442
00:21:46,000 --> 00:21:50,000
Okay, right now I don't want that to happen because we are going to focus on sequential process.

443
00:21:50,000 --> 00:21:52,000
And finally, you also have one output file.

444
00:21:52,000 --> 00:21:58,000
So all the content, all the info that we are basically getting generated, we will put us in the new

445
00:21:58,000 --> 00:22:00,000
blog post.md okay.

446
00:22:00,000 --> 00:22:05,000
So here you can see, uh, the entire thing, get the info from the YouTube channel, summarize the

447
00:22:05,000 --> 00:22:06,000
info from this.

448
00:22:07,000 --> 00:22:09,000
Uh, all the information is basically there.

449
00:22:09,000 --> 00:22:12,000
Summarize the info from YouTube channel video on the topic and.

450
00:22:15,000 --> 00:22:19,000
Create the create the content for the blog.

451
00:22:19,000 --> 00:22:20,000
I'll go and add it.

452
00:22:21,000 --> 00:22:24,000
Create the content for the blog.

453
00:22:24,000 --> 00:22:30,000
So basically whatever thing I explained my YouTube video, uh, in that way only this blog will get

454
00:22:30,000 --> 00:22:30,000
generated.

455
00:22:30,000 --> 00:22:34,000
Okay, so both these things are done perfect.

456
00:22:34,000 --> 00:22:37,000
Uh, here you can probably see I have also created my task and how easy it was.

457
00:22:37,000 --> 00:22:38,000
Right?

458
00:22:38,000 --> 00:22:43,000
So finally, to run all these things, we will also create one more file which is called as crew dot

459
00:22:43,000 --> 00:22:43,000
p y.

460
00:22:43,000 --> 00:22:46,000
And that is where for our execution will probably start.

461
00:22:46,000 --> 00:22:49,000
So here I'm going to write crew dot p y.

462
00:22:49,000 --> 00:22:50,000
Okay.

463
00:22:50,000 --> 00:22:52,000
Uh, and this is where my execution is going to start.

464
00:22:52,000 --> 00:23:02,000
So for this again uh, I will go ahead and import some important things like from crew I import crew.

465
00:23:03,000 --> 00:23:04,000
q.

466
00:23:04,000 --> 00:23:05,000
Comma process.

467
00:23:05,000 --> 00:23:08,000
Since we are going to use sequential process, that is the reason we are using this.

468
00:23:08,000 --> 00:23:13,000
Then from agents import the same thing.

469
00:23:13,000 --> 00:23:20,000
Let's see what all things we imported and from tools we have to import all these things okay.

470
00:23:20,000 --> 00:23:28,000
So from agents I don't require tool from agents I'm going to import this from from From tools or from

471
00:23:28,000 --> 00:23:29,000
task?

472
00:23:30,000 --> 00:23:32,000
From task I'm going to import.

473
00:23:33,000 --> 00:23:33,000
What?

474
00:23:33,000 --> 00:23:35,000
All things we have to import from task.

475
00:23:35,000 --> 00:23:38,000
Let's see, let's see, let's see two things.

476
00:23:38,000 --> 00:23:39,000
Researcher task.

477
00:23:42,000 --> 00:23:44,000
Researcher task.

478
00:23:44,000 --> 00:23:45,000
Research task.

479
00:23:45,000 --> 00:23:48,000
And the other task is nothing but write task.

480
00:23:48,000 --> 00:23:49,000
Okay.

481
00:23:49,000 --> 00:23:56,000
So these two are imported now Now, in order to call this in a sequential manner, we will be using

482
00:23:56,000 --> 00:23:58,000
this crew class.

483
00:23:58,000 --> 00:24:01,000
We will give our agents researcher and writer.

484
00:24:01,000 --> 00:24:09,000
So here we will go ahead and write blog researcher blog writer okay then our research task right.

485
00:24:09,000 --> 00:24:11,000
Task is already there.

486
00:24:11,000 --> 00:24:13,000
So that is all my task name that I have actually defined.

487
00:24:13,000 --> 00:24:16,000
This is the most important parameter process is equal to sequential.

488
00:24:16,000 --> 00:24:22,000
So this basically says sequential task execution is default okay I will also be showing you other sequential

489
00:24:22,000 --> 00:24:24,000
other processes as we go ahead.

490
00:24:24,000 --> 00:24:24,000
Okay.

491
00:24:24,000 --> 00:24:26,000
In the other video memory is equal to true.

492
00:24:26,000 --> 00:24:29,000
Cache is equal to true max RPM is equal to 100.

493
00:24:29,000 --> 00:24:31,000
And share underscore queue is equal to true.

494
00:24:31,000 --> 00:24:33,000
So these are some of the default parameters that is got selected.

495
00:24:34,000 --> 00:24:36,000
Now let's start the task execution.

496
00:24:36,000 --> 00:24:49,000
So here I'm going to start the task execution process with and hence feedback okay.

497
00:24:50,000 --> 00:24:57,000
So this will basically be my result is equal to Q dot kick off.

498
00:24:58,000 --> 00:25:00,000
And here I'm going to give my inputs.

499
00:25:02,000 --> 00:25:05,000
Inputs will be nothing but topic.

500
00:25:06,000 --> 00:25:11,000
Let's say one of the topic that I want to search, which is my most famous video, is nothing but between

501
00:25:11,000 --> 00:25:19,000
AI versus ML versus AI have the maximum view in this versus data science.

502
00:25:20,000 --> 00:25:26,000
Okay, finally I will go ahead and print the result.

503
00:25:27,000 --> 00:25:31,000
Okay, so this is how we kick off this entire thing.

504
00:25:31,000 --> 00:25:36,000
Now, as we kick off this crew, it is going to call this here.

505
00:25:36,000 --> 00:25:39,000
We are going to basically call this agents task process sequential.

506
00:25:39,000 --> 00:25:43,000
So automatically it will go to that particular references from that particular pages.

507
00:25:43,000 --> 00:25:46,000
And it is going to do the sequential execution.

508
00:25:46,000 --> 00:25:47,000
Now let me open my terminal.

509
00:25:47,000 --> 00:25:55,000
I hope now that my terminal is ready now let me go ahead and run Python crew.py.

510
00:25:55,000 --> 00:25:58,000
Now once I execute this, it will start the execution.

511
00:25:58,000 --> 00:26:00,000
It will start exploring.

512
00:26:00,000 --> 00:26:03,000
It will first of all query use this tool okay.

513
00:26:03,000 --> 00:26:06,000
One uh error is that open API key.

514
00:26:06,000 --> 00:26:12,000
So guys one error that we are specifically getting it is asking for a key error open AI underscore API

515
00:26:12,000 --> 00:26:13,000
underscore key.

516
00:26:13,000 --> 00:26:17,000
Now let me make you understand what exactly is this error.

517
00:26:17,000 --> 00:26:24,000
So if I probably go to the documentation page over here, one thing is that whenever we want to implement

518
00:26:24,000 --> 00:26:31,000
any application with the help of multi AI agents, we specifically require an LM model right integrated

519
00:26:31,000 --> 00:26:32,000
with our AI application.

520
00:26:32,000 --> 00:26:38,000
Because see over here if I go with respect to agents here, we are telling it to do many tasks as such.

521
00:26:38,000 --> 00:26:38,000
Right.

522
00:26:38,000 --> 00:26:43,000
Obviously tools we are specifically using, but once we get the relevant information we need to summarize

523
00:26:43,000 --> 00:26:45,000
it, we need to properly display it.

524
00:26:45,000 --> 00:26:50,000
We need to, uh, based on a simple prompt, we need to make sure that the content is generated in that

525
00:26:50,000 --> 00:26:51,000
particular way.

526
00:26:51,000 --> 00:26:58,000
So by default, if I probably consider connecting to any LMS, Crewe I offers flexibility in connecting

527
00:26:58,000 --> 00:27:03,000
to various LMS, including local model via Allama and different APIs like Azure.

528
00:27:03,000 --> 00:27:06,000
It is compatible with all Lang and LM components.

529
00:27:06,000 --> 00:27:13,000
So here you will be seeing multiple examples how you can probably connect with uh, connect the agent

530
00:27:13,000 --> 00:27:20,000
with the LM models, because at the end of the day, all the LM models needs to, uh, work along with

531
00:27:20,000 --> 00:27:23,000
the agent because most of the research work is specifically done with the agent.

532
00:27:23,000 --> 00:27:24,000
Right.

533
00:27:24,000 --> 00:27:27,000
Like a writer needs to probably write the entire block.

534
00:27:27,000 --> 00:27:30,000
So obviously an LM is required to summarize the entire block.

535
00:27:30,000 --> 00:27:36,000
So here you can see that in one of the parameter in agent, there is also something called as LLM indicates

536
00:27:36,000 --> 00:27:38,000
the large language model of the agent uses.

537
00:27:38,000 --> 00:27:42,000
By default it uses GPT form model defined in the environment using this particular model name.

538
00:27:42,000 --> 00:27:48,000
Now, uh, what I'll do in this particular video, I'll show you with the help of OpenAI API key later

539
00:27:48,000 --> 00:27:53,000
on, uh, as we go ahead, uh, we will also be discussing with respect to various open source models

540
00:27:53,000 --> 00:27:54,000
like llama.

541
00:27:54,000 --> 00:27:59,000
Then you have hugging face and all right now let's go ahead and do it with uh, OpenAI.

542
00:27:59,000 --> 00:28:04,000
So for OpenAI, what I'm actually going to do again, if you want to probably do it at any open source.

543
00:28:04,000 --> 00:28:04,000
Right.

544
00:28:04,000 --> 00:28:09,000
You can again play with this over here in the llama integration it is told you like what all keys you

545
00:28:09,000 --> 00:28:10,000
need to set right.

546
00:28:10,000 --> 00:28:12,000
And based on that you can actually work.

547
00:28:12,000 --> 00:28:15,000
And uh, how all the step by step it is given over here.

548
00:28:15,000 --> 00:28:19,000
But let me just show you with the help of, uh, OpenAI over here.

549
00:28:19,000 --> 00:28:20,000
So I will go to my agents.

550
00:28:20,000 --> 00:28:21,000
Okay.

551
00:28:21,000 --> 00:28:24,000
I'm specifically using LM is equal to LM this parameter.

552
00:28:24,000 --> 00:28:26,000
I'll try to set it up.

553
00:28:26,000 --> 00:28:26,000
Additionally.

554
00:28:26,000 --> 00:28:28,000
Now let me do one thing.

555
00:28:28,000 --> 00:28:30,000
Let me go ahead and create my LM.

556
00:28:30,000 --> 00:28:31,000
Before that.

557
00:28:31,000 --> 00:28:37,000
Uh, I have already created my environment variable dot env over here to explore the, uh, to to to

558
00:28:37,000 --> 00:28:41,000
basically explore all the to to get the OpenAI API key itself.

559
00:28:41,000 --> 00:28:47,000
So what all parameters we specifically require here you can see I will just go ahead and write import

560
00:28:47,000 --> 00:28:48,000
OS okay.

561
00:28:48,000 --> 00:28:49,000
Import OS.

562
00:28:50,000 --> 00:28:54,000
Now in this import OS what I'm doing I'm importing OpenAI API key.

563
00:28:54,000 --> 00:28:55,000
Right.

564
00:28:55,000 --> 00:29:00,000
And this API key is basically coming from my env environment variable which is having my OpenAI API

565
00:29:00,000 --> 00:29:00,000
key.

566
00:29:00,000 --> 00:29:02,000
And the second one is OpenAI underscore model name.

567
00:29:02,000 --> 00:29:05,000
Like which model I really want to use.

568
00:29:05,000 --> 00:29:05,000
Okay.

569
00:29:05,000 --> 00:29:06,000
Now now one more thing.

570
00:29:06,000 --> 00:29:11,000
In order to import all these things we have installed load underscore dot env.

571
00:29:11,000 --> 00:29:13,000
So I'm also going to call this.

572
00:29:13,000 --> 00:29:17,000
So let me go back again to my agents okay.

573
00:29:17,000 --> 00:29:26,000
And I'm going to write uh from load underscore dot env import load underscore dot env okay.

574
00:29:27,000 --> 00:29:30,000
So here you can see load underscore dot env.

575
00:29:30,000 --> 00:29:33,000
And I'm going to basically import things right.

576
00:29:33,000 --> 00:29:35,000
So this looks absolutely fine.

577
00:29:35,000 --> 00:29:38,000
Uh here sorry it should be dot env not load underscore dot env.

578
00:29:38,000 --> 00:29:43,000
Now we are uh basically we are calling all the variables from the environment variable.

579
00:29:43,000 --> 00:29:46,000
We are setting up OpenAI API key and OpenAI underscore model name.

580
00:29:46,000 --> 00:29:50,000
Now let me just go ahead and execute it because this is the thing that is required.

581
00:29:50,000 --> 00:29:53,000
And after that we will be setting with our LM in our agent.

582
00:29:53,000 --> 00:29:53,000
Right.

583
00:29:53,000 --> 00:29:55,000
So LM is equal to LM here.

584
00:29:55,000 --> 00:29:58,000
Also we can go ahead and set it in the agent LM underscore LM.

585
00:29:58,000 --> 00:30:03,000
So now I think it should start doing our process and it should start working.

586
00:30:03,000 --> 00:30:06,000
So let me quickly go ahead and write Python q.py.

587
00:30:06,000 --> 00:30:10,000
And now you will be seeing that it will start this entire process.

588
00:30:10,000 --> 00:30:11,000
Now again let me repeat it.

589
00:30:11,000 --> 00:30:12,000
Uh, in our queue.

590
00:30:12,000 --> 00:30:15,000
First of all, it will go and do the blog research for my YouTube video.

591
00:30:15,000 --> 00:30:18,000
Then it will give that information to the blog writer.

592
00:30:18,000 --> 00:30:20,000
This blog researcher is going to do the research task.

593
00:30:20,000 --> 00:30:23,000
This blog writer is going to do the right task, right?

594
00:30:23,000 --> 00:30:26,000
So here you can see now the processing of the videos has started.

595
00:30:26,000 --> 00:30:28,000
This will probably take a one minute time.

596
00:30:28,000 --> 00:30:34,000
And uh, because it is around 4 to 5 minutes, six minutes video, it will probably I have asked for

597
00:30:34,000 --> 00:30:36,000
what is AI versus ML versus data science.

598
00:30:36,000 --> 00:30:39,000
And this will probably go to my video in my YouTube channel.

599
00:30:39,000 --> 00:30:41,000
Explore the transcription.

600
00:30:41,000 --> 00:30:43,000
You know, get all the relevant information.

601
00:30:43,000 --> 00:30:49,000
And then finally, at the end of the day, uh, you will be able to see over here, we will be getting

602
00:30:49,000 --> 00:30:52,000
one new file, new blog, underscore post.md.

603
00:30:52,000 --> 00:30:53,000
So let's wait till then.

604
00:30:53,000 --> 00:30:54,000
I'll just pause the video.

605
00:30:54,000 --> 00:30:58,000
I think it will take a one minute time and then once it is executed I will show you.

606
00:30:58,000 --> 00:31:02,000
So guys, finally you can see that the entire program has got executed.

607
00:31:02,000 --> 00:31:08,000
And uh, this is the entire content of the amazing blog that is created for the video query that I had

608
00:31:08,000 --> 00:31:09,000
asked.

609
00:31:09,000 --> 00:31:09,000
Right.

610
00:31:09,000 --> 00:31:12,000
So if you also go here and see.

611
00:31:12,000 --> 00:31:14,000
So this is your new vlogging post.

612
00:31:14,000 --> 00:31:16,000
Uh, and this is the entire vlog, right?

613
00:31:16,000 --> 00:31:23,000
Whatever things I have explained in the video, everything has basically put up in a good, uh, good

614
00:31:23,000 --> 00:31:24,000
sentence in form of blog.

615
00:31:24,000 --> 00:31:29,000
Now I can directly use this, and I can even automate this if I'm probably creating a big, amazing

616
00:31:29,000 --> 00:31:30,000
end to end project.

617
00:31:30,000 --> 00:31:32,000
This will also get automated automatically.

618
00:31:32,000 --> 00:31:34,000
That entire blog may also get created.

619
00:31:34,000 --> 00:31:40,000
So here is each and every thing under the umbrella ML, DL and Data science.

620
00:31:40,000 --> 00:31:43,000
You know the words has been correct over here and this is how it has worked.

621
00:31:43,000 --> 00:31:43,000
Right.

622
00:31:43,000 --> 00:31:48,000
And if you probably see how much processing it probably took, you know, it was hardly one hour or,

623
00:31:48,000 --> 00:31:50,000
sorry, one minute of, uh, embedding.

624
00:31:50,000 --> 00:31:52,000
It has done embedding many things.

625
00:31:52,000 --> 00:31:53,000
It has done probably.

626
00:31:53,000 --> 00:31:57,000
First of all, see when we hit the query it goes in search in the YouTube channel.

627
00:31:57,000 --> 00:32:02,000
Once it gets the video, then it in removes a I mean it takes out all the transcript.

628
00:32:02,000 --> 00:32:06,000
Then finally you'll be able to see that it has been able to create this entire content.

629
00:32:06,000 --> 00:32:07,000
Right.

630
00:32:07,000 --> 00:32:13,000
So this, in short, is an amazing way of creating this creating of agent task and tools.

631
00:32:13,000 --> 00:32:17,000
Uh, and here you can also see sequential processing is basically happening.

632
00:32:17,000 --> 00:32:18,000
Researchers is doing one work.

633
00:32:18,000 --> 00:32:20,000
Two agents are also communicating.