Inspiration was that the main character's energy needed to come from a place of realness. Tired of generic chatbots that sound like corporate newsletters, the goal was to create a recommender with a personality that’s savvy, quick-witted, and friendly—basically, a digital bestie who knows movies and drops fire emojis. The inspiration was to fuse the efficiency of an LLM (to handle natural language and data parsing) with the unapologetic casualness of Gen Z culture.
What it does No cap, I'm here to solve the "What should we watch?" dilemma, but only with a curated list.
I take your vibe (a direct query, a genre, or a "rando rec" request).
I check my custom knowledge base (the movie list you gave me).
I recommend the Title, Genre, Rating, and Notes, all explained with Gen Z flair.
Most importantly, I remember your name and keep the conversation flowing!
How we built itThis whole operation is powered by a Gemini Chat Model node (or equivalent LLM node) inside an n8n workflow.
System Message is the Vibe: A detailed System Message was written to lock down my Gen Z personality, rules, and mandatory responses (like asking for the name first).
Data Injected: The custom movie list was injected directly into the System Message context so I only have to look internally.
Instruction Set: Specific rules were encoded for every scenario (Rando Rec, Genre Lock, Data Fallback, Social Questions) to ensure consistent, on-brand behavior.
Log in or sign up for Devpost to join the conversation.