← Back to home

PodcastAI

Love podcasts? Create your own with AI, monetize via ads, fair, autonomous and future-proof.

Problem Statement

This project is an integrated system for AI-driven podcast generation and monetisation, utilising a Web3 framework for financial transactions and content management. The core functionality is divided between an artificial intelligence pipeline that handles content creation and advertising, and a decentralised finance layer that manages payments and revenue distribution through smart contracts. The AI system is composed of three distinct agents designed to automate the monetisation lifecycle. The AI Content Generation Agent is responsible for creating the base podcast script, adhering to user-defined parameters such as character personalities, voice selections, topics, and overall tone. When monetisation is enabled for an episode, the AI Matching Agent is activated. This agent analyzes the podcast's content, thematic elements, and defined audience profile, comparing it against a database of active brand advertising campaigns to identify relevant and non-intrusive partnership opportunities. Once a suitable campaign is matched, the Content Generation Agent is utilised again to create a contextually appropriate advertisement script, designed to integrate naturally into the podcast's conversational flow while meeting the brand's messaging requirements. The final step in the AI pipeline is handled by the AI Verification Agent, which serves as a quality control mechanism. This agent programmatically analyses the final mixed-content episode to confirm that the advertisement was placed correctly, meets predefined quality and compliance standards, and fulfills all campaign criteria before any financial transaction is authorised.The Web3 monetization infrastructure is built around a central Solidity smart contract deployed on an Ethereum-compatible network, which governs all financial interactions between brands, creators, and the platform. For advertising revenue, brands initiate campaigns by staking their budget in PYUSD tokens directly into the smart contract, where the funds are held in escrow. These funds are only released upon successful validation from the AI Verification Agent, which confirms that ad placements have met all contractual obligations. Once verified, the smart contract automatically processes payouts, distributing 95% of the revenue to the podcast creator and retaining a 5% platform fee. This ensures a transparent and trustless payment system based on verified performance. In addition to ad revenue, the smart contract also facilitates subscription-based monetization, allowing creators to offer premium content to their audience. The system includes mechanisms for fraud detection and relies on verified view counts to ensure the integrity of all payouts. User authentication is managed entirely through Web3 wallet connectivity, utilizing wallet addresses as primary user identifiers and JWT for session management, thereby eliminating the need for traditional credentials.The platform's architecture is built on a Next.js 14 full-stack application using TypeScript for type safety. The backend logic is handled through API routes, while the frontend is rendered using React 18. Data persistence is managed by a MongoDB database with Mongoose ODM used to define structured schemas for podcasts, episodes, campaigns, and ad placement data, all indexed for performance. The audio generation component integrates with Google Gemini's text-to-speech services to produce character-specific voices, and the resulting audio files are stored for streaming. The entire system is supported by a RESTful API architecture that exposes endpoints for managing podcasts, campaigns, and AI services, with protected routes secured by wallet-based authentication middleware. Automated background jobs handle the ad verification and payout processes, ensuring the system operates efficiently without manual intervention.

Solution

PodcastAI is an autonomous podcast generation and monetization platform that merges artificial intelligence, blockchain technology, and modern web development. It enables users to create AI-driven multi-character podcasts and integrates automated, AI-validated advertising with blockchain-based payments.The platform is developed with Next.js 14 using the App Router for optimized performance, while TypeScript ensures type safety across the codebase. Tailwind CSS and Radix UI establish a cohesive design system, and Framer Motion powers advanced animations. The persistence layer uses MongoDB with Mongoose for schema flexibility, while Upstash Redis manages caching and session data, scaling seamlessly on Vercel.At its core, the AI architecture is built on Google’s Gemini models, chosen for structured JSON output and pricing efficiency. Three specialized agents coordinate content creation: the Content Generation Agent produces scripts with consistent characters and ad-ready dialogue; the Matching Agent aligns brand campaigns to relevant podcasts using scoring algorithms; and the Verification Agent validates ad placements for compliance, naturalness, and quality before payout authorization.Monetization is powered by Ethereum smart contracts implemented in Solidity and deployed with Hardhat. Wallet connections are handled by Reown AppKit, and contracts manage campaign creation, staking, payouts, and subscriptions using PYUSD stablecoin. An AI validator role enforces trustless verification, ensuring brands and creators transact without intermediaries.The AIEnhancedPipeline orchestrates the content lifecycle from concept to final audio. It generates scripts, embeds ads through natural conversation breaks, and produces multi-character audio with distinct voices. MongoDB schemas track podcasts, episodes, campaigns, budgets, and analytics, balancing creative generation with detailed advertising and payout data.Automation is handled by the AIAutomationJobs system, which manages verification, payouts, fraud detection, and analytics on scheduled intervals. Fraud detection algorithms monitor engagement for anomalies, while audit trails preserve transparency. Robust error handling ensures resilience, with fallbacks for failed AI outputs, blockchain retries, and degraded audio generation. The parseUntilJson utility salvages usable responses from malformed AI outputs.Development follows a spec-driven methodology with documentation in the .kiro/specs directory, ensuring structured feature delivery. Testing is comprehensive, spanning Jest unit tests, Hardhat smart contract checks, and AI pipeline integration tests. Performance is optimized through Next.js features, React Query caching, Redis acceleration, and batching mechanisms for AI. Blockchain interactions employ gas optimization to reduce costs.Security and compliance underpin the system, with strict smart contract access controls, wallet security, and compliance validation through the verification agent. Detailed audit logs support transparency and regulatory requirements.Overall, PodcastAI delivers a scalable, secure, and automated ecosystem where AI powers creative content, blockchain ensures trustless monetization, and modern web technologies provide seamless performance and resilience.

Hackathon

ETHGlobal New Delhi

2025

Contributors