This is a submission for the Cloudflare AI Challenge.

What I Built

UI may appear broken, but I couldn't have done better.

Technologies used in project:

FastAPi
React
Tailwind

⚠️Initially my idea was to create a simple image generator inspired by the impressive visuals of some NFTs. I wanted to throw myself deep down into the possibility of using a model to generate NFTs based on user input.

✨Additionally, I included a feature to help users improve their commands by upgrading and adding extra details into given command.
Also another feature allows users to request a quote along with an image of its author, adding a touch of authenticity.

🎈Moreover, Ask-me supports content in multiple languages by automatically translating it into English. Users can also upload audio files to receive a transcript containing all the words spoken, along with a word count.

Demo

🔥Live Demo💻

My Code

itz-Amethyst / ask-me

Your go-to for personalized NFT creation , quotes with it author image translate any context into English, can Identify the text inside an audio file (Cloudflare Hackathon)

Ask-me is a service designed to assist with various tasks, such as NFT creation, generating quotes, and translation.

🔥Live Demo💻

Technologies used in project:

FastAPi
React
Tailwind

UI may appear broken, but I couldn't have done better.

⚠️Initially my idea was to create a simple image generator inspired by the impressive visuals of some NFTs. I wanted to explore the possibility of using a model to generate NFTs based on user input.

✨Additionally, I included a feature to help users improve their commands by providing feedback and suggestions. Another feature allows users to request a quote along with an image of its author, adding a touch of authenticity.

NFT_Model = "stable-diffusion-xl-lightning"
Chat_Model

…

View on GitHub

Journey

Given that this was my first touch with workers and AI models, the circumstances were undeniably challenging. In particular, the pressure to complete the task within just 1/5 day it did cost sacrificing sleep

If I had more time, there are a few things I'd like to add or improve:

Enhance the user experience on the frontend, especially in terms of design. Making it more visually appealing can make a big difference.
Implement a voice recording feature for whispers. It could add a personal touch and make interactions more engaging.
Enable a feature to save pictures to the D1 database, associating them with user IP and browser information as identifiers.
Find a solution to deploy Python workers along with their requirements on Cloudflare. This could streamline the deployment process and make scaling easier.

Multiple Models and/or Triple Task Types

NFT_Model = "stable-diffusion-xl-lightning"
Chat_Model = "llama-2-7b-chat-fp16"
Improve_NFT_Command_Model = "hermes-2-pro-mistral-7b"
Quote_Model = "mistral-7b-instruct-v0.2"
Translate_Model = "m2m100-1.2b"
Audio_Model = "whisper"