logo

Content Catalyst: AI-Powered Platform-Native Content Generation

One Sentence Pitch

Content Catalyst is an AI content generator that translates a business brief into platform-native, tonally consistent content ideas instantly, eliminating creative block and ensuring brand cohesion.

Inspiration

The problem is the marketing waste cycle: Startups spend 15-25% of annual revenue but burn creative hours struggling to ideate. We saw that unknown brands are achieving viral growth—going from zero to 600K followers with millions of impressions—using low-cost, short-form video. The challenge is turning that potential into consistent, brand-aligned content.

What It Does

Our application takes a user's business description, a target platform (e.g., TikTok, LinkedIn), and a desired tone (Witty, Professional), and instantly generates five ready-to-use content ideas. Each idea is custom-formatted for the channel, including suggested visual hooks and scripts, making content strategy actionable in seconds.

How We Built It

The Flask backend acts as a secure API gateway, utilizing the Anthropic Python SDK. The core LLM is Claude 3.5 Sonnet, which executes a zero-shot, multi-step system prompt to enforce strict persona, tone, and platform-specific formatting before returning the structured output to the minimalist HTML/JavaScript frontend. All sensitive API keys are secured via environment variables.

Challenges We Ran Into

Our biggest challenges were timeboxing and communication in the tight 24-hour window, requiring aggressive task delegation. Technically, we faced significant hurdles in prompt engineering to coerce the LLM into returning a reliable, machine-readable format (strict JSON structure) while simultaneously maintaining creative freedom for the content.

Accomplishments That We're Proud Of

Functional Deployment: Delivering a fully working, end-to-end deployed web application in under 24 hours.

Prompt Engineering Sophistication: Successfully enforcing platform-specific formatting (e.g., vertical video script structure) using Claude's reasoning capabilities.

Architecture: Setting up a robust, scalable Python/Flask production environment (using requirements.txt, Procfile, and Gunicorn).

What We Learned

Project Scaffolding: Gained critical experience quickly starting a full-stack Python project from scratch to deployment (using requirements.txt, Procfile, and Gunicorn).

LLM API Management: Learned best practices for secure API key handling and the importance of server-side rate limiting for consumption models.

Teamwork: Optimized communication and rapid task division in a compressed timeframe.