AI image generation system through dialogue

I created an intelligent system that allows generating and improving images in simple terms. The user does not need to know prompt engineering — the system itself transforms the idea into an accurate technical request and produces ready-made images.

How it works:
the user describes the idea in any language;
the AI agent translates and optimizes the request for the generation model;
remembers previous edits (style, light, mood) and takes them into account;
generates 3 vertical images (1080×1920) per request;
shows the result immediately in the chat without unnecessary links.

Result for the user:
quick creation of visual content without technical knowledge;
convenient refinement through dialogue;
stable quality for Stories, Shorts, and advertising;
low cost of generation.

Tools: n8n, Claude 4.5 Sonnet, Flux 2 Dev (ModelsLab), JavaScript.