VibeArchitect

Inspiration

In the current creator economy, the difference between a viral hit and a buried post often comes down to "vibe-matching"—the subtle art of picking the right song for a specific location and a high-conversion thumbnail. We noticed that even the most talented creators spend upwards of two hours per post on this manual guesswork. We were inspired to build a tool that didn't just automate the creative process, but architected it using objective data.

What it does

VibeArchitect is an agentic creative suite that automates trending audio selection and cinematic thumbnail generation based on location data. It uses the Presaige API to audit and iteratively refine visuals until they meet a 9.0+ viral readiness score.

How we built it

We built a React/Vite frontend and Node.js/Express backend, orchestrating a pipeline of Gemini for reasoning and Flux for image generation. The core logic involves a closed-loop feedback system that programmatically adjusts prompts based on Presaige's technical recommendations.

Challenges we ran into

We faced significant API rate limits during the 24-hour sprint, necessitating a custom "Waterfall" fallback system to maintain 100% uptime. Mapping subjective "vibes" to objective geolocation data also required extensive prompt engineering to ensure authentic results.

Accomplishments that we're proud of

We are incredibly proud of our resilient "Cascading AI Waterfall," which guarantees application stability despite third-party service outages. Additionally, turning a static diagnostic API into a programmatic, self-correcting creative director was a major technical win.

What we learned

We learned that AI is most effective as a feedback loop, where diagnostic data is used to steer creative generation rather than just label it. This project proved that engineering resilience is just as critical as the AI's "intelligence" in building production-ready agentic workflows.

What's next for VibeArchitect

We plan to implement an inbuilt photo editor for manual adjustments and a feature for users to upload their own images for auditing. Further down the roadmap, we aim to expand into frame-by-frame video auditing and direct automated social media posting.

Built With

• express.js
• flux.1-(black-forest-labs)
• font-awesome
• framer-motion
• google-gemini-flash
• google-gemini-flash-lite
• google-gemini-pro
• javascript-(es6+)
• lucide-react
• pollinations.ai
• presaige-ai-(recommendations-api)
• presaige-ai-(score-api)
• react
• stable-diffusion-(sd-xl)
• vanilla-css
• vite