Nexsar

Our project is a blockchain-based platform that seamlessly integrates decentralized AI with Human Intelligence Tasks (HITs), revolutionizing the way content is created, curated, and analyzed for public sentiment. Designed for distributors such as businesses, content creators, and influencers, the platform automates the creation of high-quality content while engaging decentralized networks of users for feedback, content evaluation, and sentiment analysis. By harnessing the power of AI models like DALL-E and OpenAI, content is generated based on specific user briefs, offering tailored and dynamic outputs. Blockchain integration ensures secure, tamper-proof operations, providing transparency and accountability in content management.

Distributors access the platform through the LIT Protocol, which guarantees decentralized access management and secure data handling via encrypted LIT Wallets. AI-generated content is then reviewed by a decentralized network of workers (users) who evaluate, vote, and provide feedback through a transparent, blockchain-based voting system. Workers are incentivized with platform tokens, ensuring active participation and trust in the curation process.

At the core of this system is the Galadriel on-chain AI agent, which dynamically manages AI-driven content generation. This agent offers customizable AI models for both image and text creation, allowing for flexible, high-quality outputs that meet the exact needs of distributors. The blockchain-based voting and transparent verification processes ensure the platform maintains integrity and decentralization throughout the workflow. The LIT Protocol adds another layer of privacy and security, encrypting sensitive user and content data while streamlining decentralized operations.

The platform also features NFT minting for valuable content, granting distributors full ownership of their digital assets and enabling monetization via an integrated NFT marketplace. This system not only facilitates decentralized content creation and curation but also supports decentralized content moderation, giving workers a voice in ensuring content quality and compliance. By combining the latest in AI automation, blockchain security, and decentralized collaboration, the platform delivers a seamless, scalable, and trustworthy ecosystem for content management in the Web 3.0 era.

[ Detailed Deck link Is given here : https://www.canva.com/design/DAGP7vpZguQ/VpBKSeIyX1_5XkpRAy41bA/edit?utm_content=DAGP7vpZguQ&utm_campaign=designshare&utm_medium=link2&utm_source=sharebutton ]

Our project combines cutting-edge blockchain technology, AI, and decentralized tools to create a robust content generation and curation platform.

Frontend Tech Stack:

1. Lit Protocol:

• Session-Based Wallet Signing : LIT Protocol generates session keys that allow users to securely access the platform without the need to manually sign every transaction. When a distributor or worker logs in, LIT Protocol assigns a Public Key, Private Key, and a Session Key, eliminating the need for repeated manual wallet signing. This streamlines actions like content generation, task management, and post verification, reducing workflow friction and enhancing automation.

• Role-Based Access Control : LIT Protocol manages access via role-based session keys, ensuring secure, specific permissions for distributors and workers. A distributor can access content creation tools and NFT minting, while workers access content verification tasks. This prevents unauthorized actions and ensures data security throughout the platform.

• Encrypted Content and Feedback Transmission : LIT Protocol provides public-private key encryption to secure content and feedback data. For example, when workers review AI-generated content, details are encrypted to ensure feedback remains unbiased and protected. Similarly, worker feedback is encrypted and securely transmitted, guaranteeing data integrity and privacy.

• Off-Chain Computation for Task Metrics : LIT Protocol facilitates off-chain computation for votes, feedback, and verification metrics. For instance, worker votes and feedback on AI-generated content are processed off-chain, reducing the need for costly on-chain transactions. This makes task management and post-verification more scalable and efficient.

• Automated NFT Minting and Transaction Signing : LIT Protocol automates the NFT minting and signing process. Once content is finalized, it is automatically minted as an NFT, and the process is securely handled by LIT Protocol without manual intervention. This allows distributors to easily tokenize their content for trade or ownership, ensuring security and efficiency in transactions.

• Broadcasting of Task Results and Metrics : LIT Protocol supports off-chain broadcasting of task results, enabling decentralized processing of worker feedback, votes, and task metrics. This eliminates the need for on-chain interactions while maintaining transparency, reducing costs, and increasing speed in task reporting and completion.

• Data Privacy for Sensitive Features : LIT Protocol encrypts sensitive data such as budget and post frequency settings. Distributors can modify and access this encrypted data, ensuring that only authorized users can decrypt and view sensitive business information. This guarantees data privacy while maintaining flexibility in content adjustments.

• Transparent Task Execution and Verification : LIT Protocol automates the signing of task completion and verification processes. Once workers complete a task, the system automatically signs off the task and sends it to the distributor for approval, eliminating manual intervention. This improves transparency, trust, and security in task workflows.

• Multi-Node Distributed Key Management : LIT Protocol ensures high security through multi-node distributed key management. Session keys, private keys, and access controls are securely distributed across multiple nodes, protecting transactions, content access, and verifications from potential threats. This decentralized key management enhances platform reliability and data protection.

1. Galadriel AI:

• AI-Driven Content Creation Using Galadriel AI : Galadriel AI dynamically selects models like DALL-E or OpenAI’s GPT to generate high-quality content, such as text-to-image or image-to-image outputs. The distributor provides input briefs including preferences like style, target audience, and tone. Based on these inputs, Galadriel AI generates personalized content tailored to the distributor’s specifications. This automation streamlines content creation, reducing the manual workload for distributors and increasing content quality.

• Dynamic Model Selection : Galadriel AI operates as an on-chain AI agent that intelligently selects the most suitable model for each task, such as DALL-E for generating images from text descriptions or GPT for generating textual content. This dynamic selection ensures that the content generated matches the specific requirements and complexities of the distributor’s brief. By integrating with multiple AI models, the platform ensures versatility and precision in content creation. We have integrated a text-to-image AI agent and can do the text-to-text as well as image-to-image conversion.

• AI-Generated Content Curation by Workers : All AI-generated content is posted securely to the worker dashboard, where workers can curate the content by reviewing multiple versions and selecting the best fit. Workers also provide feedback on the quality, style, and accuracy of the generated content. This decentralized curation system, combined with Galadriel AI’s high-quality outputs, enhances the overall content validation process.

• Text-to-Image and Image-to-Image Generation : Galadriel AI supports advanced features like text-to-image and image-to-image generation can also be done. Distributors can input text descriptions or existing images to generate new visuals, which are refined based on distributor requirements. The AI ensures that each generated image meets the quality standards and aligns with the branding or messaging goals of the distributor.

• Feedback-Driven AI Refinement : Workers provide detailed feedback on each AI-generated post. Galadriel AI uses this feedback to refine its future content generation, ensuring continuous improvement. For example, workers may suggest changes to tone, style, or visual elements, and Galadriel AI dynamically adjusts its output based on this input, improving the quality of the generated content over time.

• NFT Creation from AI-Generated Content : After content is curated and verified by workers, the platform automatically converts the best AI-generated outputs into NFTs. Galadriel AI plays a pivotal role in generating high-quality digital assets that can be tokenized. These NFTs can then be sold or traded in Web 3.0 marketplaces, providing distributors with ownership over their unique digital assets.

1. Ether.js & WAGMI: These libraries handle blockchain interactions. From wallet connections to transaction management, they streamline decentralized operations, making the platform both user-friendly and secure.

2. Pimlico (Paymaster and Bundler): The Paymaster and Bundler features of Pimlico play a vital role in optimizing gas efficiency across the platform.

• Paymaster allows the platform to cover or defer gas fees for users, particularly for microtransactions, making it affordable for workers to participate in tasks without worrying about high gas costs. This ensures that all interactions, such as voting or content verification, can be completed without workers or distributors being burdened by unpredictable transaction fees.

• Bundler aggregates multiple small transactions (such as worker votes) into a single, larger transaction, optimizing gas costs and reducing blockchain congestion. This feature is critical for maintaining the economic viability of microtask completion on the platform.

1. React Frontend : The frontend, built with React, ensures that distributors and workers have a smooth and intuitive user experience. • Metamask Integration: Users can easily connect their Metamask wallets for decentralized login and transaction signing, which is crucial for interacting with blockchain-based content generation and moderation tasks.

• Content Brief Creation: Distributors can input their content briefs and customize their requirements (e.g., style, audience, frequency) for Galadriel AI to generate high-quality posts.

• AI-Based Content Generation: The platform displays AI-generated posts for workers to evaluate. Workers can vote on the best options, with all votes being transparently recorded on-chain.

• Automated Transactions with LIT PKP: Using LIT PKP (Programmable Key Pair), the platform encrypts the entire system and automates transaction signing, allowing users to interact with the platform without manually signing every transaction. This reduces friction in the user experience while maintaining security.

The React frontend enables smooth user experiences, with functionalities like Metamask integration, content brief creation, and AI-based content generation. Workers evaluate posts and vote on them, with their votes recorded transparently on the blockchain. The entire system is encrypted, and transactions are signed automatically without user intervention via LIT PKP (Programmable Key Pair).

Backend Architecture:

1. Prisma (Database Schema Design): We use Prisma to define and maintain the database schema, ensuring a well-structured, normalized database that handles the vast amounts of content data, user profiles, and reward distributions. Prisma ensures that all operations on the database remain consistent, supporting seamless content and task management.

2. PostgreSQL Database: As a highly reliable and consistent relational database, PostgreSQL serves as the core database for efficiently storing and retrieving data. This includes content briefs, user data, AI-generated content, task assignments, and reward distribution details.

3. Atomic Transactions: We implement atomic transactions to guarantee smooth and error-free operations. For example, when workers vote on content or receive rewards, these actions are processed as atomic transactions, ensuring that they either fully succeed or fail and roll back in case of an error, preserving the integrity of the system.

4. Cron Jobs for Automated Content Generation: Cron jobs are used to trigger automated content generation based on the frequency set by distributors. Galadriel AI generates content at set intervals, and LIT Protocol automates the signing and pushes these tasks to the blockchain without requiring manual intervention, ensuring timely content delivery.

Smart Contract Integration:

1. Turks.sol (Decentralized Voting Contract): The Turks.sol smart contract enables decentralized voting for workers evaluating AI-generated content. Distributors create posts, and workers vote on various options, which are then recorded transparently on the blockchain. This contract ensures the integrity and transparency of the voting process, with all data stored immutably on-chain.

2. DalleNft.sol (NFT Minting Contract): The DalleNft.sol contract integrates with the teeML oracle to generate and mint NFTs based on user prompts and Galadriel AI-generated content. Using the ERC721 standard, the platform converts AI-generated content into NFTs, allowing distributors to monetize their work through blockchain-based ownership. These NFTs can then be traded on decentralized marketplaces.

3. Gas Optimization with Batched Transactions: To further minimize gas fees, we utilize batching mechanisms where multiple micro-transactions (e.g., worker votes, task completions) are grouped together and pushed to the blockchain in a single transaction. This dramatically reduces the cost of on-chain interactions, making the platform more scalable and cost-efficient.

In summary, we’ve built an intricate decentralized content platform with seamless integration between AI, blockchain, and encryption technologies. The use of Lit Protocol for security, Galadriel for AI, and atomic transactions in the backend ensures that our platform operates efficiently, securely, and scalably, with low gas costs and high transparency.