Inspiration
Small businesses and creators lack time and resources to plan engaging, multi-platform social campaigns. We imagined an AI “genie” that translates product details into ready-to-post content in seconds.
What it does
MarketGenie ingests your campaign info—budget, product, audience—and uses Claude to auto-generate post titles, descriptions, schedules and formats for Twitter, Instagram, LinkedIn, etc., then saves them for review or direct publishing.
How we built it
• Frontend: Next.js + Auth0 for SSO and user profiles
• Backend: Express.js + PostgreSQL on AWS RDS for campaigns & assets
• AI: Anthropic Claude via @anthropic-ai/sdk to craft JSON post arrays
Challenges we ran into
• Handling database connectivity and SSL setup on RDS
• Parsing strict JSON from LLM responses and validating the schema
• Synchronizing Auth0 sessions between Next.js and Express, fixing CORS, and also setting up the X-ai
Accomplishments that we're proud of
• End-to-end flow from user signup to campaign creation, asset generation, and listing
• Reliable Claude integration that produces consistent, valid JSON
• A polished, responsive UI in minutes with Chakra’s theming
What we learned
• Real-world Express/Postgres setup on AWS RDS, including migrations
• Robust error handling for LLM outputs and TypeScript event typing
• Best practices for Auth0 in Next.js + separate API, plus CORS & security
What's next for MarketGenie: Your AI Marketing Agent
• Integrate direct posting to social APIs (Twitter, Facebook, LinkedIn)
• Enable media uploads and AI-assisted image/video creation
• Add analytics dashboard with engagement metrics
• Build scheduling workflows and reminders (n8n / DuploCloud MCP)
• Expand to more sponsors’ tools (Apify scraping, Senso.ai insights)
Built With
- amazon-web-services
- anthropic
- chatgpt
- claude
- express.js
- javascript
- llms
- minmax
- nextjs
- node.js
- python
- react
- x-api
Log in or sign up for Devpost to join the conversation.