Speaker 1: In this video I'm going to show you the correct way to use the newly released amazing DeepSeek R1 reasoning model inside NADN. The reason I'm going to say correct way is a lot of people are actually showing you the wrong way to use this inside NADN because they have an older version of NADN and it's not going to work if you have the updated version. I'm going to show you step by step what I'm talking about. But just quickly let's talk about what DeepSeek R1 is and why this is such an incredible release in the ecosystem of AI. So first of all AI reasoning model with DeepSeek R1 are extremely important in the AI ecosystem because they're designed to think through problems, make decisions and solve complex challenges just like we do. Now for that reason they are a lot slower compared to normal GPT-4s of the world because of the fact that they require thinking and complex reasoning and therefore they actually use a lot more computing power and they're very very expensive. So for example if you were to compare the GPT-4-01 model or the OpenAI-01 model right the input token is $15 per million token the output is $60 per million token and that's why if you have a chat GPT subscription they only allow you I think 10 per week or something similar to that. And then also if you're using it through the API it becomes very very expensive. But the DeepSeek R1 model which again is on par as far as the performance with the GPT or the OpenAI-01 model is extremely cheap. So for context it's $0.55 per million token and $2.19 per million output token. This is incredibly important and incredibly cheap because of the fact that for bigger and more complex problems these models are amazing. So for example for solving things in the medical industry, research or things like enabling robots to navigate the world or just for personal reasons creating personalized recommendations to make our lives easier, these models are perfect. Therefore if you want to use these models inside End2End to make your workflows a lot smarter and if you're creating automations that are for more complex workflows that require reasoning this is the perfect way to use this inside End2End. So let's go ahead and I'm going to show you step by step how to use this newly released DeepSeek R1 inside End2End. If you're new to the channel my name is Zubair. My YouTube channel, my school community is all about building amazing AI agents with no code tools like End2End. Our community is incredible. They are full of passionate people who are there to learn and collaborate with each other. I'm going to put the link in the description so make sure you join because it's an amazing space for you to get involved because again the best way to learn in my opinion is through a community and this community is amazing because it's not only me but Alex and Alexandra who are all three of us are End2End ambassadors have combined our efforts to create the best possible community for everybody to learn. We have a great classroom section where you can grab any of these AI automations, the workflows you can directly import it from right here, the JSON template and you can use all of the automations. On top of that we have amazing classroom section where regardless of your background we have introductory series, we have exclusive deep dive topics on voice AI agents, factor databases, AI agents and much much more. We also have an amazing community automation workflows. Our community is always building new workflows, new ways to use End2End on their personal lives, on their businesses so you can grab all of this right here. We also have a calendar section, we have daily calls, tech support, we have live build every Thursday. Alexandra builds a new project so you can join the call and follow her along as she builds a new project live build. If you have any questions obviously you can always post them in the community, somebody or myself will get back to you within a few hours and also if you miss any of the calls we have our call recording section in the classroom where you can access everything from there. Again if you're serious about learning AI, how to build AI agents with no code tool like End2End, this is the perfect space to be. I'm going to put the link in the description, make sure you join and I'll see you there. Alright let's get back to the video. If you have updated your End2End instance, the user interface is completely different. That's why I'm saying that a lot of people that are showing you on YouTube how to use this inside End2End with the AI agents through the open AI's chat model is actually wrong. So make sure you are in the latest model and the way to check that is you see right here if there's any updates available, it will show you update one or two and then you'll go to admin panel and update and then once you come back so what you need to do is, again this is just for test purposes, you can always add this for example when I created Jarvis. Again you can add that here for making these more complex workflows but I'm just going to show you just on the chat model how to use this quickly so that way you can be on your own as far as like using this for further workflows. So let's go ahead and actually just do a chat. I'm just going to add a AI agent here. So I'm going to go to advanced AI, click on AI agent. Alright so now let's go ahead and back out. So I'm going to show you how to attach this. Whoops. Alright so go ahead and click on chat model here and you're going to use open AI chat model but I'm going to show you how to use open router instead. So you're going to come to your credentials obviously if you're using open AI, this is going to have your open AI account and then all of the models here that are loading is going to be GPT models of course. So the way to add this DeepSeq R1 quickly on your workflow is going to the credentials here. You will create a brand new credentials and on the API key, so this is the big difference here. Before on the older versions of NNN, this base URL right here was actually outside of the credentials. It was right here where you had the option to add this base URL right here but now it's gone as you can see. So if you are seeing the base URL here, it means you have the older version of NNN. So make sure you update that because that's very very important because now like I said on the new credentials, the base URL is sitting right here. So the way to do this is first of all you have to have an open router account. So you would sign in on your account. You would head over to your account here and just click on keys. You'll create a new API key and then you'll come back here and all you have to do is just paste the API key right here. So once you do that, it's very important for you to change the base URL. So then you would go over to your, you can just go to DeepSeq R1 or any other model and if you just scroll down in the bottom, so right here, this is where you just you can use this open AI or open router API directly. All you have to do is just copy the URL here right here and if you copy this, you'll come back and you'd paste it on your base URL because I've already created several accounts. I'm going to show you actually. So if I go to my open router account here, all right there you go. So right so this is my API key that I posted and this is where I pasted that base URL and again same thing. It's just openrouter.ai slash API slash v1 and again that's exactly what this is right here. Okay so that's what you do. You make sure you paste that there and then you'll go ahead and get your connection with test post say successfully. So now again before the base URL was here, now it's inside. The problem is now unfortunately this drop down is completely gone. So with this new user interface, this drop down where before you had the option to click on this drop down and all of the models that was inside open router were available, now it's not. So the workaround for this is to actually click on expression and here's what we need to do. We need to manually copy this. So if we go back to our model here, so your whatever LLM you're using or whatever AI model you're using, when you go inside that model in the bottom you can see right here, you can just copy this right here or just click this little button. This will copy this model name here and now all you have to do is come back and literally paste this here and now you should be you will be able to use this and all the models that are inside open router. So let's go ahead and don't worry about this little sign here and I'm going to show you the proof to actually you can use this right here because we can go ahead and check out our activity. So if you come over to your account here and click on activity, so this is where you'll see all of the different models or you have interacted with through the open router API and the last one I used was the Google Gemini Flash. So I'm going to show you exactly now how to use this DeepSync R1. We're going to test it out and I'll come back and show you that it's actually worked. Alright so let's go back. I'm going to go ahead, you don't need to do memory or tools and stuff like that because like I said I'm just testing this. You can always just replace your normal workflow by adding the open chat model and going through step just that I walk you through right here to add it into your more complex workflow. So let's go ahead and click on chat and let's go ahead and test this out. So I'm going to ask it a complex question. What is the meaning of life? Right so now it's going to go ahead and process it and if you did everything right this is going to reach out to that DeepSync model and it's going to process our request and as you can see it does take longer because like I said this model is designed for complex reasoning and it will give you a more a longer answer and a complex answer as well. There you go say the meaning of life is deeply personal philosophical blah blah blah right. So now let's go ahead and check to see if indeed we use this model. So if I go back to my open router and I just refresh the page right now again like I said the latest model last time I used was Gemini Flash. So if I refresh the page there you go perfect right this is DeepSync R1. I just used it and as you can see a little it's a little more expensive but believe me it's nothing compared to compared to the OpenAI 01 model and you can see right here on the right hand side the provider DeepSync and we did everything correctly. So again make sure you're using this method if you have the latest version of NNN installed because if you use the previous method or some of these guys who are showing you on YouTube it's not going to work. It's going to completely confuse you and I actually had done the same mistake previously when I did my video when I showed how to use the DeepSync V3 model. I was using the previous model of NNN or the previous version of NNN and it didn't look like this and a lot of people got confused. So in my school community I made an update to make sure people are aware of that. I hope you found this video helpful. If you like it please like and subscribe and if you're serious about learning how to use NNN in your complex workflows and business and personal life make sure you join the school community. Thanks for watching and I'll see you in the next one.