Introduction
Syndi was born out of our curiosity to find out how quickly we can implement a chatbot from scratch without using any existing builder platforms. Along the way, we thought the large machine learning models released recently such as Stable diffusion and GPT-J would be a great addition to this experiment and thereby we can build something that has real use.
Â
Functionalities
Syndi is a chatbot where it has many features including
- Stable diffusion game
- Text to 3D
- Text to Art
- Brainstorm
- Create AD copy
- Generate product description.
Â
With Stable diffusion, you can guess the correct prompt to make an image.
- Sci-fi
- Fantasy worlds
- Cartoons
- Pets
- Paintings
Â
You can create stunning 3D illustrations with your words using stable diffusion
Â
You can create stunning paintings with plain words using stable diffusion.
Â
Syndi (chatbot) is powered by GPT-J where you can create an AD copy for your product and services as well as product descriptions
What technologies did we use?
The entire implementation took about a day. We went with the tech stack that is scalable at the same time quick to implement.
For frontend
- React js
- ChatUI
For backend
- Node js
- GPT-J - for NLP functionalities including Chat, AD copy generation, Product description generation
- Stable diffusion - We use a search engine that is powered by Stable diffusion for listing images generated for functionalities including Text to 3D, Text to Art, and Stable diffusion game
Â