VIDEO TRANSCRIPTION
AICodeKing
DeepSeek-V3 Agents : This FULLY FREE AI Agent can DO ANYTHING!
Description generated by Skrybot
A new AI model, DeepSeek V3, is available for free via Hyperbolic's API. Users can easily create AI agents using VectorShift, an AI automation platform. By creating workflows in VectorShift, users can connect data sources, build custom workflows, and develop AI agents for various tasks. The process involves dragging blocks to create workflows, connecting inputs to AI models like LLMs and DeepSeek, and adding integrations like Notion for context. Users can test and deploy their workflows as chatbots, automations, or APIs. DeepSeek offers cost-effective performance for various automation tasks.
Transcription generated by Skrybot
Hi, welcome to another video. So, DeepSeek V3 has been recently launched, and it's a really good model, especially because it's fully open weights, and their API is dirt cheap, which is pretty great to see. Not just that, you can currently use the DeepSeq V3 model for free via Hyperbolic. Yes, although you won't see DeepSeq in Hyperbolic's model section yet, they are providing the model for free through their API. It's a little unstable for now, and until it gets added to the platform officially, it seems to be free, which is great. So, that's super great. And many people were asking me how they can create agents with the new DeepSeek V3 model to accomplish tasks that are needed in any custom workflow you'd like for it to do.
Well, that's what I'm going to show you today. Most people want to make these AI agents easily as well, without messing around with anything locally and stuff like that. So, to keep everything simple, we'll be using VectorShift. If you don't know about VectorShift, I have some other great videos about it as well. But in general terms, it's an AI automation platform that allows you to make workflows where you can connect any data source to AI, create custom workflows, and build AI agents that you need. Like, if you want an AI agent that can summarize your Google Calendar events, you can do that. Or, if you want an AI agent that reads your Notion notes, and summarizes them for you, it can do that as well.
You can also put these workflows in a chatbot, and you can also access these workflows via APIs. Not just that, you can also automate them, like making your workflows trigger when you send a message in Discord, receive an email, or something like that. So, that's super cool as well. Now let me tell you how you can configure it all, and create agents with the new DeepSeek V3 platform. So, just get yourself signed up here. It also has a free tier, which is good. Anyway, once you're here, you can just go to Pipelines. Pipelines are the major stuff, and everything else is dependent on this. Now, just hit the new button.
and it will open up this page where you can select from some pre-built workflows or you can just hit this and it will open up a new canvas here you can drag in the blocks you want to connect and create your own workflow so let's say that we are creating a simple chatbot Well, in that case, we'll first need to drag in an input here. This will allow the user to give input. Now we can connect this input to something, like an LLM or a knowledge base, where you can easily make a user input query, a knowledge base, and give the required context from that knowledge base. Or, you can also connect it to some other integration. But because we are creating a simple chatbot, let's add an LLM here.
Now, in LLMs, you have multiple options like OpenAI, Anthropic, perplexity, or something like that. Now, to use DeepSeq, we can just drag in the custom one here and enter the DeepSeq or Hyperbolic API base, URL, model name, and API key based on what you want to use. Once that's done, you can connect your input to it and also enter the system prompt. Once done, you can just drag in the output and then connect it to the LLM. Now you have a basic workflow that takes input, runs it through an LLM, and gives you the output. Now, you can add whatever you want to this workflow as well. Let's say you want the AI agent here to have the context of your Notion notes.
You can just drag the Notion integration in and connect it in the workflow as you'd like, and your LLM will have access to that, which is great. You can even add search or some other LLM and route some requests to that LLM based on if-else logic. This allows you to delegate specific tasks to another LLM or do some other kind of flow, which is all super great. You can also test the pipelines as you make them by clicking this button and just entering the input and testing it through here. This will allow you to check everything before you deploy it as an API. or chatbot or anything like that. So that's also cool.
Once you have everything figured out, you can just click the deploy changes option and it will save your workflow. Then you can just export it to be an automation, chatbot, search or form, which is pretty great. Let's select chatbot. Now once you select it, you can just enter the chatbot name and then go to the next page. Here, you can configure stuff like how you want the chatbot to look and things like that. Once done, you can just deploy the chatbot, integrate it anywhere, or use it as an API in your own applications. Or you can also just chat with the chatbot here, which is also great to see. So, that's how you can configure an AI agent with DeepSeq easily via VectorShift.
Apart from this, you can also create automations with it. Automations are basically a way to trigger the workflow you've made via something that happens in another application. For example, if you want your workflow to be triggered when you send a message in some Slack channel, you can do that quite easily here, which is pretty great. There are also options to add knowledge to your AI agents. You can just upload the files that your AI agent can reference, and take in your custom knowledge base context, which is also pretty great. You can also create voicebots, bulk jobs, portals, evaluations. and stuff like that, which is great. So, that's how you can easily create AI agents with DeepSeek's new model.
It's super cool to use for these kinds of workflow automations, because it is super cheap, and it allows you to get the same performance as Claude in these tasks for super cheap, which is great. You can also add search, any kind of apps knowledge base, and stuff, which makes it so cool. It's all in a drag and drop interface that allows you to customize it in any way you want and use it all for free via something like Hyperbolic with VectorShift, and it will all be free. Overall, it's pretty cool. Anyway, let me know your thoughts in the comments. If you liked this video, consider donating to my channel through the SuperThanks option below, or you can also consider becoming a member. by clicking the join button.
Also, give this video a thumbs up and subscribe to my channel. I'll see you in the next video. Till then, bye. .
S K R Y B O T
Skrybot. 2023, Video transcriptions from YouTube
By visiting or using our website, you agree that our website or the websites of our partners may use cookies to store information for the purpose of delivering better, faster, and more secure services, as well as for marketing purposes.