InfoCast

Users often encounter disjointed interactions across multiple platforms. The limited word count on social media profiles also restricts the amount of information users can provide, hindering the ability to tailor interactions and services to individual preferences effectively.

Furthermore, FAQs on websites are somewhat ambiguous and challenging to navigate. Fortunately, many companies and teams utilizesome form of AI or GPT-chatbot on their websites to address these some of these issues. However, transitioning from social media to external websites can/would disrupt the user experience.

To finese our way out of these challenges, we've developed the InfoCast frame. One that allows users to engage in seamless conversations directly with the platform, meeting urgencies and eliminating the need for redirection.

By converting SVG prompts and responses to PNG format and storing them on Pinata, InfoCast is able to deliver an In-line chatbot on Frames.

Ceramic's ComposeDB serves as the decentralized database for storing chat history. Each thread of conversations is initiated by minting an NFT to the prompter and using the token-balance access control on Lit protocol to encrypt the convos stored on ComposeDB.

The NFT tokens are minted are linked to a users Farcaster ID through the Privy's embedded wallet generated behind the scenes for user upon signing in

We fetched an already stored vector embeddings and added user prompts before querying OpenAI's LLM for a natural response