Using Google Gemini API with No Code in AI Workflows

hey Builders there's been a big advancement in AI that you need to know about Google have just released their successor to Bard called Google Gemini this isn't just another interation but rather a gigantic leap forward with benchmarks either matching or exceeding what we've come to expect from chat gbt 4 but what does this mean for us well a lot actually because our build chip workflows they're about to get a whole lot smarter Google Gemini isn't just a chatbot it's an advanced large language model trained on massive amounts of D what more it's multimodal meaning it can input and output both text and images this means Nuance text summaries image classification and creative content Generation all at your fingertips Gemini can tackle tedious text based tasks analyze product reviews localize your content for Global reach or quickly brainstorm fresh campaign ideas all within your existing buildit workflows so let me introduce you to the Google Gemini AI node you'll find the Google Gemini nodes under Gemini in the node Explorer there are a couple of specific use cases here for example we've got a node that counts the number of tokens in a prompt so you can keep track of your spend in the API but for reference Google tell us in their documentation that for Gemini models a token is equivalent to about four characters so 100 tokens are about 60 to 80 English words we're going to focus on the multimodal Node today so on the right we have our node instructions you start by getting a Google AI API key and you get that from AI Studio create an API key and add it into the secrets manager so it's securely stored it comes conveniently in under the secrets menu here next we choose our model at time of recording Google had released Gemini Pro and Gemini Pro Vision publicly available as part of their apis there is also an image generation model called image Gen 2 which is available to Google trusted testers and Google have just released Gemini 1.5 which you can go on a wait list for you can be sure that will update our template nodes as these and more new models get added into Google's API let's do a quick example using Gemini Pro let's create a build ship workflow that looks a little something like this it'll use the pre-built Bing search node to search the web for the latest news articles on a topic we'll feed this prompt in as part of our API call so under query we click the little plus call this prompt and then we can feed that in here query prompt now this being search node out puts adjacent object so before we pass it on to a node that requires a string we're going to need to stringify that Jon and buildship has a pre-built node that will do that for you you simply feed in the Bing search results now for our Gemini node here I've created a prompt which gives it adjacent object of news articles on a specific topic from the past 24 hours and asks us to tell us what the key points from that news wrap up are so hit ship and let's test it out we're going to send the query prompt in as a search for new news articles from the past 24 hours about say low Cod app development and here we have our key points from the last 24 hours of news it's given us some expert stats and it's even pulled out a gardner prediction now our Gemini Pro Vision model Works in much the same way but we have a couple of extra parameters to play with and that's to make it really easy for you to upload the image that you're going to send over to Gemini I found this excellent looking slice of cake on unsplash and I want to get Gemini to not only tell me what's in the image but I want it to give me a recipe of the food it finds within that image you just click the plus icon and upload your image so this uses the power of build chip storage as a back end you just need to tell up the file extension all right we'll hit ship and test this out our prompt will be what food is in this image provide the recipe and here you have it it's identified the lady finger biscuits as well as the mus Pony cream and giving you the full instructions for the recipe so as you can see there's a huge amount of power in this note stay tuned for our next video where we dive deeper into specific Gemini features and tutorials to get you up and running don't forget to like share and subscribe for more updates and until then Happy building

Channel: BuildShip

Views: 1,702

Keywords: API builder, no-code backend, low-code backend, automation builder, AI powered workflow builder, visual backend, backend builder, no code google cloud, low code google cloud, no code AI workflows

Id: Lnp5G7K-Ar4

Length: 4min 29sec (269 seconds)

Published: Fri Mar 01 2024