Thank you my friend 🙂 Do subscribe if you haven't already & share with others too!

Hi ! Yeah of course. You could create a scroll/carousel of featured apps for example and on click use the zoom navigation transition to show an app details screen.

Its true that it's sort of 'limited' in a sense. I also love MatchedGeometry & use it extensively, however I've often run into issues when using it inside NavigationStack for instance. I'm still playing around with it for now.

@@swiftodyssey thanks for sharing new features

@@RussChannel13 you're welcome, happy coding !

You're welcome ! 🙂

Hi @Navy-729 ! It seems the issue occurs in the latest version of the SDK v0.5.4 as of this writing while the tutorial used v0.4.4 when published. I've fixed the issue & pushed changes to use the latest version of the SDK here: github.com/anupdsouza/ios-gemini-chat/tree/multimodal Please take the latest source code from here. Hope this helps, & do consider subscribing to my channel if you found the content useful !

@@swiftodyssey thanks a lot!!!

@@Navy-729 you're welcome!

@@swiftodyssey One more thing: while the code works and I tested it, the model does not keep track of the context of the previous messages and cannot build off of prior responses. I was thinking that I would append the history of each message prior to the user's message, but is there a more efficient way to do this?

@@Navy-729 Yes, you can. You need to initialise a `Chat` instance for that. Take a look at the text-only chat with Gemini video here where I make use of it: kzfaq.info/get/bejne/jLVjjKp_l7vZeJc.html At the time of that video, the SDK wouldn't support text & media based interactions but more recently, it seems that the SDK is able to manage it. You may need to play around a bit to figure things out. Here's how Google does it using the vertex sdk but the code is almost exactly the same firebase.google.com/docs/vertex-ai/chat?platform=ios. Also take a look at my vertex ai video on how to send photo & video media from this point onwards kzfaq.info/get/bejne/l6t9erh3nLnRoqM.html. Do note that the context for the model is limited to a certain extent depending on the gemini model you use such as pro or flash. I might update this sample when I get some time. In case I do, I'll drop a line here. Hope this helps!

Hey! That would require working with Firebase Storage. Check out the documentation here: firebase.google.com/docs/storage/ios/upload-files . Consider subscribing to my channel if you haven't 🙂

@@swiftodyssey Thanks , I have subscribed already

Hi ! The gemini SDK will work the same way with UIKit since the AI request code is not UIKit dependent. Once you get the AI response, you can display it in a UITextView or UILabel connected to, for instance an IBOutlet. Is there any particular challenge that you are facing? Let me know, and do subscribe to the channel if you haven't already :)

Hi! There are a couple of reasons for this. #1. If there was an issue getting a story from Gemini API which could be due to incorrect api key, api response limits or incorrect json data OR #2. If there was an error retrieving images for the story received in step #1. This can happen if you dont have the a proper Open AI secret key, there is a rate limit or you don't have enough credits for Dall E to generate images. I'd advise setting a breakpoint in the code and checking the execution step by step.

@@swiftodyssey Hello, -[RTIInputSystemClient remoteTextInputSessionWithID:performInputOperation:] perform input operation requires a valid sessionID. inputModality = Keyboard, inputOperation = <null selector>, customInfoType = UIEmojiSearchOperations im getting this error i search this error nothin help can you help please!

That is unrelated to the issue. Set a breakpoint at line #49 here github.com/anupdsouza/ios-gemini-storybook/blob/main/Storybook/Storybook/Library.swift and check the response you receive from Gemini. If all is well, you should see the story printed via the print statement at line #63. Next, see if you get an error printing at line #73 & #88. That will execute if there was an issue creating images.

@@swiftodyssey thank you it’s showing green thread 1 breakpoint 49,73 what does it mean

Hello! All the good ones are paid & lack a free trial.

Thank you! Do consider subscribing to the channel !

Hi. Something is definitely broken with the SDK when using images. I've reported an issue on the SDK page on Github. Hopefully someone from the dev team can identify the problem. I will post an update here as & when I receive it. Thanks for bringing attention to this, also don't forget to subscribe to the channel :)

Hi, I tested with both the 0.4.4 version of the SDK used in this video as well as the latest 0.4.8 and I no longer see error 500 in the responses. The issue was temporarily occurring due to service related issues which are now resolved: github.com/google/generative-ai-swift/issues/122

Glad it helped!

Hi ! So there are 2 parts to this. The first is a config plist file where the key is stored and the second part is the APIKey.swift helper file which helps to read the key from the plist. I have shown this in detail in my starter video here kzfaq.info/get/bejne/jpOqppCoxpzOYmw.htmlsi=vFy-0KAK1e4lP5ix & I'd highly recommend that you begin with that video first to understand the setup better. Alternatively, you can also download the code github.com/anupdsouza/ios-gemini-chat/tree/multimodal and copy your api key in the Generative AI plist file and play around. Hope this helps 🙂

Thank you so much buddy. Appreciated :)@@swiftodyssey

@@user-eu6ui7wq7t You're welcome! Do leave a like & consider subscribing if you found the videos helpful. 🙂

Thank you! Make sure your target device is iOS Simulator/Device. Also I'd highly suggest checking out a copy of the code from the repository here github.com/anupdsouza/ios-gemini-ai-horoscope and giving it a go. If you still run into issues, create a new issue in my repository with a link to your project for me to take a look at it github.com/anupdsouza/ios-gemini-ai-horoscope/issues

Thank you! Do give the video a like & subscribe if you haven't already :)

:P Already did that @@swiftodyssey

Thank you! Appreciate it 🙂

Hi Ion! It’s great to hear that you enjoyed my videos so thank you for your support :) As for the number of subscribers, it’s probably because (a) my channel is fairly new & (b) the content is geared more towards stuff that I'm currently interested in, so the video topics can be quite niche. Take for example my GTA VI logo drawing video or the Scratch Card Effect video which are not quite common; unlike many other talented iOS educators with content that is focussed on more commonly researched topics. Having said that, I’m always looking for ways to improve my content and grow my channel. If you have any suggestions or ideas for future videos, please let me know! 😊

Thanks! Leave a like & consider subscribing if you liked this content 🙂

Hope the video helped you out! Do like the video if it did and consider subscribing to the channel

Glad you liked it !

@@swiftodyssey Can you show how we add image in this chat please

@@funstudiovlog8148 the gemini-pro model does not support text and image support at this time. For images you will need to use the gemini-pro-vision model. I will be uploading a video on that soon. Stay tuned!

@funstudiovlog8148 New video with image is up! kzfaq.info/get/bejne/armVipSpmdOad5c.html

It is just one of the ways by which you could store your API_KEY. Refer Google's documentation here: ai.google.dev/tutorials/swift_quickstart#secure-api-key . You could store and read your API key any way you want, just ensure that you don't expose it in public source control.

Leave a like & consider subscribing if you liked this content 🙂

Glad you like it!