
In this video, I'll introduce you to VectorShift, a powerful no-code AI automation platform, and show you how to use its functionalities for various use cases, including agents, chatbots, and enterprise-ready assistants. You'll learn to create accounts, build pipelines, and integrate different nodes with LLMs from providers like OpenAI, Anthropic, and Gemini. I'll guide you through integrating Google Docs to create dynamic workflows, using email automations with Gmail, and configuring type forms and white-labeled chatbots. Watch as I build a complete pipeline from scratch, demonstrating VectorShift's power and flexibility. Links: https://vectorshift.ai 00:00 Introduction to VectorShift 00:30 Creating Your Free Account 00:38 Understanding Pipelines 01:22 Leveraging Integrations 02:09 Data Handling and API Connections 04:32 Building a Chatbot from Scratch 06:43 Introduction to LLM Integration 06:58 Exploring Google Docs Integration 07:43 Configuring Google Docs with LLM 08:48 Querying and Testing with Google Docs 10:05 Advanced Features and Use Cases 11:40 Creating Automations with LLM 13:49 Conclusion and Next Steps
--- type: transcript date: 2024-10-21 youtube_id: mGD1pRMv3T8 --- # Transcript: VectorShift: The No-Code Platform for AI Agents, Assistants, Chatbots and Automations in this video I'm going to be showing you Vector shift which is a noode AI automation platform there are a ton of different use cases in here from Agents to automations to chatbots all the way through to assistants that are Enterprise ready these can be for professional use cases these can be personal use cases things like content generation internal company search co-pilots in this video I'm going to walk you through exactly on how you can get started with building pipelines to build out whether it's automations chatbots or really whatever you can imagine on their platform so first thing that you can do is you can make a free account on Vector shift there's a ton of capabilities that you'll see on the left hand sidebar here first I'm going to show you pipeline pipelines are arguably at the heart of vector shift and what you can do if I just click this simple chatbot template at its core what it allows you to do is you have these nodes that you can drag onto the screen here let's say you have a knowledge base that you want to use you can drag this here and then you can ultimately wire it up the cool thing with this is you can leverage a ton of different lmus anything from open AI anthropic llama Google AWS models perplexity and you can just start building you don't even necessarily need to do this in a successive order let's say we want to use perplexity because maybe we want to query something like the internet but for instance let's just say I want the workflow to trigger a draft within my email I can do that as well now one of the most powerful features of this are the Integrations here there's everything from GitHub through to slack through the Google Suite whether it's Gmail gcal Google Docs you can use things like Google Docs as knowledge sources if you'd like or you can have them perform certain actions you can also use something like Gmail say if I want to trigger it and make a draft email or actually send an email out you can build out workflows from that you you can also even reference things like what you have in notion or whether you want to use pine cone or postgress there's really quite a comprehensive Suite in terms of different capabilities and Integrations that you can leverage within what ever you want to build this could be a chatbot use case this could be an Automation and you can build all of the different whether it's automations chatbots all within this interface here in terms of the types of data that you can load in you can have structured or unstructured data anything from YouTube to Wikipedia csvs you can even query external apis if you'd like you can really do just about anything and that's something that I really want to emphasize with this platform is this isn't necessarily for people that don't know how to code this can absolutely be for develop velers even though you don't need to write out any code you can build it out in a way where you can leverage it within your application and that's another thing that I wanted to touch on is depending on what you build with the pipelines you can also connect to an external API as well now there is the ability to even generate images so say if you want to generate an image with Dolly 3 you can have that be an output you do have the ability to have text to speech from things like 11 Labs or you can even Leverage The Vision capability of some of the popular models whether it's gbd4 anthropic series of models or the Gemini series of models as well now in this example you can really put anything within the system message now to deploy it once you've set up whatever you want to set up you do have the ability to run this just like you would a chat boot so if I just say something like hello world and I submit that you'll see it work through the node so you can see the runtime you can see the number of AI credits that it costs and then you'll see the response there and the nice thing with this is once you're ready to export this so you can export it as an automation or a chatbot or even a form if you'd like so let's say I want this to be a chatbot and I want to call it example bot I can just create the chat bot and then I have this interface here where I can just start to configure it so let's say I don't want voice messages I can do that I can change the welcome header so instead of it saying Vector shift you can have this be whatever you want I can put in developers digust you can also shrink and change the image and you can do this All without codes if you go to export once you're done you do have the ability to include it within an iframe or you could just add it within the script tags now you also have the ability to connect this from slack alternatively if you want to just run this from the environment of wherever you're building out your application you can include it by just copying whether it's the python JavaScript or just take the curl request and this is a good example where you could put it in chat GPT and really convert it to whatever language you might be working in so the one thing to note is you will have to get an API key if you're going the route of the API export but you'll be able to find that right within the interface so let's go back and let's actually build one out completely from scratch we'll just call this developers digest example So within here so the first node that we're going to drag onto the screen is just the input so you can think of this as what might be a user message you also have the ability to change the type so say if you want to have multiple inputs you can do that as well so say if you have a form or something that might have multiple inputs depending on the fields that there are you can do that upload audio or files as well and in this case we're just going to have one input from there what we're going to do is we're just going to select the llm that we're going to be using in this case I'm going to drag over anthropic you have the option to select the model here and I'm just going to click Sonet 3.5 here once you have that what you can do is if you want to reference the variable in the other nodes that are within your graph here is you can just use double curly braces and then you can specify the name of the field that you want to reference in this case let's say I want to reference input one field I can just reference that here and you can also name it something and give the prompt some further context you could say something like here is my input now using XML tags can be useful as well if you just say like user input here and then user input that could be a strategy you don't necessarily need to do that that has been an anecdote that I've heard that some people have had success with just something to consider now the other thing with this is you can just specify what you want it to be you could say something like this is the user's question or you could say something like the user query to wire this up is we can just connect this input here and we'll see that now we have this blue dot here that connects to this input now the other thing that you might potentially want to do is you might want to name this something a bit more descriptive so you could name this something like user question if you'd like and then also within here you're creating the query you could also specify something like user question now let's just go back to the general here and we'll just put in an output so this is going to be very similar to the first example that I showed you now within the output could just go ahead and run it at this point this is going to be a working chat bot and here if I just deploy this I go and run this I'll submit that we see that it will go through the LL there we go we have the working example now that's not really exactly an interesting use case right this is essentially something like akin to a GPT wrapper or in this case an anthropic wrapper I'm going to show you the Google Docs integration and the reason why I think this is an interesting use case is you could create this graph and then you could hand this off to someone and so long as they know something like Google Docs is they could feed in the context that would ultimately power knowledge base of this particular workflow that could be a chatbot or it could be used in an automation or what have you all that you need to do is work through the signup flow to begin to add in the integration once you're all logged in what you can do here is is you have the ability to search the docs you can search through all of the different Google Docs that you have or you can load the particular context of one particular Google doc let's just say I want to read a Google doc you can go and you can click configure once your account is synced as soon as you create new files on Google Docs you'll be able to easily integrate them right within the interface here in this case I'm just going to select one file that I have here and we're going to save that configuration now now that I have this I can create a new variable and if I just call this something like Google doc as soon as I have that I'll have the ability to create this connection here now the one thing that is nice with this is you can have warnings now if you have a node that could run into a potential issue it's going to show this warning state for you where this can be an issue is let's say you have a Google doc that is particularly long you might need to add in a subsequent node that will actually Index this document and be able to perform some sort of similarity search or rag functionality to be able to retrieve all of the contents that are within this document but for this sake the document that I'm going to be showing you is relatively short so in terms of the context length it should be able to plug in to the llm that we're using and just about any llm given that is a relatively short document in this case I just have some generated data with some specific values that we're going to be using to query and test this once we finish building out the pipeline now that we have the user question as well as the contact here just to reiterate we're saying you're a helpful assistant that always answers the user's question based on the context answer in a conversational manner what we can do here is I'll just put this side by side here and now that we have this all wired up and we'll just use this for reference so let's just say what is the reader for the publication we'll just query that and what's going to be need about this is all of a sudden now we're going to be referencing this live working Google doc now if we just look through this example we see a very detailed example and the thing with this is since we're passing in the entire context of this entire document it shouldn't be a big surprise that that we do get a response but you can see how this could be potentially helpful right it has all of the very nuanced metrics that I have within here the target audience 78% professional developers 22% students and hobbyists we can go down here we can see that it's grown significantly Etc and so on and so forth like I said this isn't true data this is just some generated data that I used I generated it with Claud before I built this example where this is very powerful is if I just go and paste this out with completely different information so again here's another CLA generated set of information without even touching anything what is the Quant script framework and I send in that query we see the processing request and then in short order we have our answer here so here we go I'll scroll to the top here so Quantum script was released September 17th 2023 and I'm not going to read through this but you start to get an idea right because as soon as someone has access to just this Google doc or whatever the integration is all of a sudden it becomes very easy to maintain right you could do an absolute ton with this like tomorrow you could literally have an agency where you make these chat Bots for automations and you just say Quantum script chatbot I can create that chatbot you can config fig this so you can just change this out to whatever you want Quantum script change out the logo right you could change out the powered by and pretty much everything within here now I do want to show you what this looks like so you can open this up in a full chatbot you have a URL to Vector shift where you'll be able to deploy this or alternatively you can put that script within wherever you're going to be deploying this application so you could really put this wherever you want wherever your application lives you can create a new domain or what have you and you can pretty much white label basically everything within here and there is also some other really nice features like you can add in password protection if I just say test for instance and you refresh your application you'll be able to see that now you have to authenticate another cool thing with this that I'm going to be playing around with personally is the ability to leverage this within Gmail there is a great example in here of how you can generate a draft response and what you can do with this just to point you in another Direction on the platform here is you can create these automations let's just say I want want to create an email automation you can go over to automations click new select Gmail you can select the event that you want to respond to let's say when a user gets a new email for instance you can go ahead and decide what trigger you want to have you can go ahead and select your pipeline whatever it might be and you can begin to map out the different fields here now the other great thing with this is you can create automation let's just say I wanted to add a type form to my website for collaboration requests or something like that I could walk through the steps on type form set up a type form and then once it's all set up I'd have the ability to map the different fields to the particular pipeline workflow that I want to have run this could be as soon as you get that type form request that it sends to a slack Channel or creates a Discord message or it could be something like drafting a reply or anything really the thing to remember with this is since you can now leverage llms within this is you can really steer the direction for instance maybe instead of spending so much time combing through emails that are more or less going to be the same types of reply and might just be different variations of common requests let's just say for my channel and I have a request for a collaboration generally speaking the inputs are more or less going to be roughly the same but there's going to be some information sharing back and forth you could create a workflow and create a draft or you could actually even trigger to send an email and something like that and it could really be anything because by being able to Leverage The llm you can start to steer it in a way that has your tone your style as well as some potential information on whatever you want to share for instance within here let's just say you have the particular message that you get from a user you could have a work through this node and then from this node you could specify how you want to have the llm respond you can attach that as the body I just wanted to give you a really quick overview how you can get started with leveraging Vector shift and in particular building out a no code chatbot and if you want to see more of these no code examples let me know in the comments below but otherwise that's it for this video if you found this video useful please like comment share and subscribe otherwise until the next one
Weekly deep dives on AI agents, coding tools, and building with LLMs - delivered to your inbox.
Free forever. No spam.
Subscribe FreeNew tutorials, open-source projects, and deep dives on coding agents - delivered weekly.
Technical content at the intersection of AI and development. Building with AI agents, Claude Code, and modern dev tools - then showing you exactly how it works.