-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathoriginal_transcript.txt
More file actions
47 lines (47 loc) · 3.41 KB
/
original_transcript.txt
File metadata and controls
47 lines (47 loc) · 3.41 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Timestamp Text
--------- ----
00:00.400 If you're building AI agents, you've probably heard about MCP or Model Context Protocol.
00:05.240 MCP is a new open source standard to connect your agents to data sources such as databases or APIs.
00:11.679 MCP consists of multiple components.
00:13.939 The most important ones are the host, the client, and the server.
00:17.219 So let's break it down.
00:19.530 At the very top you would have your MCP host.
00:22.570 Your MCP host will include an MCP client.
00:25.910 And it could also include multiple clients.
00:28.370 The MCP host could be an application such as a chat app.
00:33.630 It could also be a code assistant in your IDE, and much more.
00:38.520 The MCP host will connect to an MCP server.
00:41.600 It can actually connect to multiple MCP servers as well.
00:48.799 It doesn't matter how many MCP servers you connect to your MCP host or client.
00:53.189 The MCP host and servers will connect over each other through the MCP protocol.
00:57.630 The MCP protocol is a transport layer in the middle.
01:04.930 Whenever your MCP host or client needs a tool, it's going to connect to the MCP server.
01:09.290 The MCP server will then connect to, for example, a database.
01:12.909 And it doesn't matter if this is a relational database or a NoSQL database.
01:17.069 It could also connect to APIs.
01:20.500 And also the API standard doesn't really matter.
01:23.120 Finally, it could also connect to data sources such as a local file type or maybe code.
01:30.590 This is especially useful when you're building something like a code assistant in your IDE.
01:36.579 Let's look at an example of how to use MCP in practice.
01:40.120 We still have the three components.
01:41.599 We would have our MCP host and client,
01:46.019 of course, we also have a large language model,
01:53.170 and finally, we have our MCP servers,
01:56.030 and these could be multiple MCP servers or just a single one.
02:03.430 Let's assume our MCP client and host is a chat app,
02:07.022 and you ask a question such as, what is the weather like in a certain location or how many customers do I have?
02:13.289 The MCP host will need to retrieve tools from the MCP server.
02:18.719 The MCP server will then conclude and tell which tools are available.
02:23.319 From the MCP host, you would then have to connect to the large language model
02:26.633 and send over your question plus the available tools.
02:30.960 If all is well, the LLM will reply and tell you which tools to use.
02:38.280 Once the MCP host and client knows which tools to use, it knows which MCP servers to call.
02:43.599 So when it calls the MCP server in order to get a tool result,
02:48.846 the MCP server will be responsible for executing something that goes to a database, to an API, or a local piece of code,
02:59.069 and of course, there could be subsequent calls to MCP servers.
03:02.889 The MCP server will apply with a response, which you can send back to the LLM.
03:07.849 And finally, you should be able to get your final answer based on the question that you asked in the chat application.
03:15.060 If you are building agents, I'd really advise you to look at MCP protocol.
03:18.800 The MCP protocol is a new standard which will help you to connect your data sources via MCP server to any agent.
03:25.280 Even though you might not be building agents, your client might be building agents.
03:29.439 And if you enjoyed this video, make sure to like and subscribe.