.png)
.png)
Hey everyone,
Hoping to provide some insight on how I was able to build a "custom" LLM that leverages all no-code tools such as bubble and I used Flowise, (drag n' drop flow builder) as the backend.
https://github.com/FlowiseAI/Flowise
In my case I made it a simple PDF uploader utilizing the Conversational Retrieval QA chain and got it working directly with bubble.io. I think this really opens up the flood gates, especially people without a technical background.
though it wasn't completely plug n' play, I did have to set it up and deploy it on render.com which was pretty straight forward. They had all this documented in the readme. However you don't get much bandwidth with render's free hosting tier. But for testing purposes that should be sufficient enough.
After I created the simple PDF uploader chain, I simply leveraged the given APIs and plugged it directly into bubble's API connector solution. They don't really have any authentication as of now so I wouldn't say it's a scalable robust solution but it gets you more than halfway there.
Now combining this backend with the power of bubble, you can do about everything. I have a video going in-depth on how I did this step-by-step if you guys are interested you can check it out below.
https://www.youtube.com/watch?v=O7NhQGu1m6c
From the shortfalls of this open-source solution, I'm building a fully robust application of my own that can directly integrate into no-code tools with bubble.io being the first one. It will be incredibly easy to use your own company data for LLMs. And you wouldn't even need to set up any of the APIs. I'm f*cking stoked to get this released!
If you guys are interested in this solution, we're giving away 50% off exclusively for the first 100 users that signup to our no-code plug n' play LLM builder (direct plugin with bubble.io)
https://41v8knyf54x.typeform.com/to/xLERO0kb?typeform-source=www.youtube.com
