Integrating Replicate with Bubble.io | Three different image generation models
Вставка
- Опубліковано 6 лип 2024
- If you have an amazing AI startup idea that you'd like to build and test, then this video is for you.
Replicate.com provides thousands of AI models in many categories.
In this video, you will explore how to integrate Bubble.io with Replicate.
You could follow steps in this tutorial to build your very own AI SaaS.
We'll cover three different image generation models with different inputs, outputs and API calls.
He shows how to authenticate.
⏰ Chapters:
00:00 - Intro
00:36 - API Tokens/Replicate API Calls.
1:05 - Bubble API Connector
1:17 - Authentication
2:11 - List Predictions
3:23 - Create a Prediction
4:50 - Input
5:22 - Image Editing Model
5:44 - Text Model
6:34 - Configuring API Calls in Bubble.io
10:00 - Where to find recent API Calls
10:30 - Input/Output differences with AI Models
11:47 - Stable Diffusion Image Generation
14:48 - Recap
Subscribe to my channel / @zubair_nocode
Also checkout
Twitter / zubair_nocode
Linkedin / zubairlk
Website
www.azkytech.com/
Build apps without code using Bubble. Sign up today
bubble.pxf.io/21RdqA
#Nocode #ai #stablediffusion #texttoimage
Thanks for the tutorial!
Really great tutorial! Looking forward for more ai bubble integration videos.
Thanks, will do!
very helpful content
Nice, good efforts, I appreciate your teaching, would love to see the video generation session too
It is already live. Have you seen ua-cam.com/video/0Zd05bzKjJQ/v-deo.html&lc=UgzNbjl_WG_llqo_YaJ4AaABAg
Very good video I would like to see how u design the front end off the web app using the api
Very great video, I wanted to use Replicate to upscale an image in bubble, how could I upload and image and send it to replicate ?
I understand how you can receive the output but not how you can pass the image input to the replicate model.
GOAT
Very cool video. Thank you taking your time to explain this to us. I was wondering if you could perhaps also explain some of the front end design for complete beginners to bubble in a similar tutorial (implementing replicate with bubble io). I saw in your "Text to Video app" video you emphasized a bit on the front end but I still don't understand how you made the area which shows the results. Thanks!
Well explained and worth watching! Thanks for the time you give , but is there any way to make the model’s id dynamic, so that one does mot need to re-initialise the call.
Too tricky as the response is very dynamic. Each model has its own input json structure and its own output json structure.
Even if we make the model ID dynamic, the input/output json formats will get complicated
How can i show the "output" image result at the bubble frontend?
👏🏼
Thanks for the tutorial! I'm wondering how many workload units it would take to generate an image. Meaning, a user sends a text prompt to replicate, and replicate returns an image. How many Bubble workload units would that consume? I'm wondering if it's even worth doing this considering the high WU cost.
Wouldn't worry about WU now tbh. Not much
I am struggling using the webhooks aswell as integrating strip, Could you touch on these subjects in the next video?
Yes webhooks can be challenging. I've noted it down to make a separate video on replicate and webhooks.
@@zubair_nocode thank you very much :)
@@zubair_nocode That would be great indeed cause I am struggling to have the web hook and in general the Bubble workflow work...
I'm facing the same problem
Here is the video on webhooks: ua-cam.com/video/sEb0eGNgopM/v-deo.htmlsi=LtchWxOlCFePmY7V
Brilliant! I guess in Bubble you have to create a workflow where you create the prediction and then get the prediction from the id of the create prediction. Is that correct? But how do you know when the create prediction is completed go get the prediction?
You can use webhooks for that.
@@zubair_nocode thanks again for those tutorials.
Do you have the tutorial for image to image generation?
I have one which uses a control net model that has image to image. It is uploading