In product wise speaking, this is just an POC demo for how the user can input a simple prompt and get a use IG post back (around 100 words). On the other hand, in tech wise, this build is supported with complete serverless stack at the infrastructure level so it can handle requests at scale. Security measures are not implemented yet, so service will be shut down immediately if unusual activities were discovered.
About
Doing copywriting for social media is a tedious task for many SMEs. BrandingCopy.AI would be the first AI helper that combine researching, generating and censoring content all in one to provide viral yet solid content for businesses. Project Scope: This project is a functional POC Demo. The main purpose is to put different parts together and gives a functional output within a limited budget and time. Key Features to ship: 1) Vector database: vector database is needed to increase the limit of how many contexts to be feed into the AI model 2) Web-scraping: leverage search engine to scrape relevant content in order to increase copywriting quality 3) AWS Bedrock: leverage AWS AI Models to generate copywriting 4) AI content censor: iterate with AI to improve copywriting quality. Especially, a) copywriting should not be detected as AI generated b) copywriting needs to be highly correlated to user inputs. c) copywriting tone and structure needs to be highly aligned to Instagram post references.