Dark

DARK brings AI-powered movie discovery into Slack. Ask Nyx for personalized recommendations, explain your Movie DNA, collaborate on shared watchlists, and discover what to watch without leaving Slack

Comment

Inspiration

Movie discovery has become increasingly frustrating despite the rapid growth of streaming platforms. Users often spend more time searching for something to watch than actually enjoying a movie. Recommendations are repetitive, content is fragmented across multiple streaming providers, and most platforms operate as “black boxes,” offering little explanation for why certain movies are suggested. Additionally, planning a movie night with friends or family usually involves switching between streaming services, messaging apps, and search engines, making the experience inefficient and disconnected.

DARK was inspired by the idea of creating an intelligent, explainable, and collaborative movie intelligence platform. Rather than simply recommending movies, we wanted to build an AI-powered assistant that understands user preferences, explains recommendations transparently, visualizes viewing behavior through Movie DNA and Taste Personas, and enables seamless collaboration directly through both the web platform and Slack. The vision was to transform movie discovery from passive scrolling into an engaging, personalized, and AI-driven experience.

What it does

DARK is an AI-powered Movie Intelligence Platform that combines intelligent recommendations, explainable AI, behavioral analytics, and collaboration into a single unified experience.

The platform enables users to search movies, discover personalized recommendations, view streaming provider availability, watch official trailers, and organize movies into custom watchlist folders. Its collaborative co-viewing feature allows users to merge watchlists and taste profiles, making it easier for friends or families to decide what to watch together.

At the core of the platform is Nyx, an AI-powered operating system designed specifically for DARK. Nyx understands user intent, answers platform-related questions, explains Movie DNA and Taste Personas, provides personalized recommendations, and assists users through natural language interactions. The platform also extends this experience into Slack, allowing users to search movies, retrieve recommendations, and collaborate directly within their workspace.

DARK also provides an analytics dashboard that visualizes viewing behavior through Movie DNA, Taste Personas, genre distributions, provider usage, and AI-generated summaries, making recommendation decisions transparent and understandable.

How we built it

DARK was developed as a full-stack web application using a modern, modular architecture focused on scalability, maintainability, and security.

The frontend was built with HTML, CSS, and JavaScript, featuring a fully responsive user interface optimized for both desktop and mobile devices. The backend was developed using Node.js and Express.js, with MongoDB serving as the primary database for user management, watchlists, analytics, and AI-related data.

Movie information, trailers, and streaming provider availability are retrieved using the TMDB API. Authentication is implemented using JWT-based secure login mechanisms.

The AI capabilities are powered by Google Gemini through a centralized AI Gateway. Instead of acting as a generic chatbot, Nyx uses intent detection, context building, structured prompt generation, caching, and optimized API orchestration to provide contextual and personalized responses while minimizing latency and API costs.

The platform also incorporates collaborative Slack integration, enabling users to interact with Nyx through Slack using secure webhook verification, OAuth authentication, and Block Kit interfaces.

To improve reliability and performance, DARK implements intelligent caching, request optimization, rate limiting, secure API communication, and modular architecture principles inspired by Hexagonal Architecture, allowing the system to remain scalable and extensible.

Challenges we ran into

Building DARK involved several technical and architectural challenges.

One of the biggest challenges was designing Nyx to behave as a platform-specific AI assistant rather than a generic chatbot. This required developing an intent detection pipeline, contextual reasoning system, and secure AI orchestration that could provide relevant responses without exposing internal implementation details or sensitive information.

Integrating multiple external services, including TMDB, Google Gemini, and Slack, while maintaining consistent user experiences and minimizing latency required significant optimization.

Creating explainable recommendation systems such as Movie DNA and Taste Personas required designing behavioral analytics that continuously adapt to changing user activity while remaining intuitive and meaningful.

Performance optimization was another major challenge. Managing AI request costs, implementing intelligent caching, reducing duplicate API requests, and optimizing modal loading times required careful architectural decisions.

Finally, balancing feature-rich functionality with a clean and responsive user experience while ensuring security, scalability, and maintainability was an ongoing challenge throughout development.

Accomplishments that we're proud of

Some of our key accomplishments include:

  • Developing Nyx, a platform-specific AI operating system capable of understanding user intent and providing contextual assistance.
  • Designing Movie DNA and Taste Personas to make recommendation systems transparent and explainable.
  • Successfully integrating Google Gemini into a structured AI pipeline that prioritizes security, performance, and personalization.
  • Extending the AI experience into Slack, enabling collaborative movie discovery directly within workspaces.
  • Building a responsive, production-ready full-stack application with modern UI/UX principles.
  • Implementing secure authentication, rate limiting, intelligent caching, and modular architecture for improved reliability and scalability.
  • Creating collaborative watchlists and co-viewing capabilities that enhance group decision-making for movie discovery.

What we learned

Developing DARK provided valuable experience in building production-oriented AI applications that combine multiple technologies into a cohesive platform.

We gained a deeper understanding of integrating Large Language Models into real-world applications while controlling costs, improving response quality, and maintaining security.

The project reinforced the importance of modular software architecture, performance optimization, caching strategies, and scalable backend design.

We also learned how to design AI systems that are explainable, user-focused, and tightly integrated with application functionality rather than functioning as standalone chatbots.

Working with multiple APIs, secure authentication, Slack integrations, responsive frontend development, deployment pipelines, and AI orchestration significantly strengthened our full-stack development skills and our understanding of enterprise software engineering practices.

What's next for Dark

Our long-term vision is to evolve DARK into a comprehensive AI-powered entertainment intelligence platform.

Future development will focus on expanding Nyx into a more capable AI operating system with richer reasoning abilities, proactive recommendations, and deeper personalization. We also plan to enhance collaborative features by supporting group recommendation sessions, real-time synchronization, and shared AI conversations.

Additional goals include introducing multilingual support, advanced behavioral analytics, voice-based interactions, cross-platform integrations, personalized notification systems, and deeper integrations with workplace collaboration tools.

On the technical side, we aim to further optimize AI inference, strengthen caching strategies, improve scalability for larger user bases, and continue enhancing security and deployment automation.

Ultimately, our vision is to transform DARK from an AI movie recommendation platform into a complete, explainable, and collaborative entertainment intelligence ecosystem that helps users discover, understand, and enjoy content more intelligently.

Built With

Share this project:

Updates