venus-assistant

As a user, I want to see an immersive Home page featuring a dynamic animated galaxy map where each star represents a feature (Ask Questions, Image Prompts, Daily Tips), an interactive floating avatar of a smart playful girl that animates on interaction (winks, smiles, shrugs), micro-interactions like ripple effects on hover, and smooth zoom-in transitions on star click. Implement based on the existing Home (v3) JSX design. Entry point to the app; leads to Chat page on interaction.

As a user, I want a backend API endpoint that accepts image generation requests and returns detailed, realistic, English image prompt descriptions ready for use in AI image generators. The API should leverage the configured AI models (Google Nano Banana for image generation) via LiteLLM routing.

As a user, I want the backend to intelligently route AI requests using LiteLLM across GPT 5.4 (user-friendly responses), Claude 4.6 Opus (academic/coding), Gemini 3.1 Pro (friendly responses), and Google Nano Banana (image generation). Configure Langchain for conversation chaining and context management. Ensure fallback and routing logic is robust.

As a user, I want the entire application to default to Hindi language and Indian locale settings (IST timezone, INR currency formatting). Configure i18n for React frontend and ensure all AI responses are delivered in natural Hindi. Apply locale settings across backend services as well.

As a user, I want my conversation history to be persisted in the database so I can continue previous conversations. Backend API should support storing and retrieving chat sessions per user, using MySQL/MariaDB with Alembic migrations.

As a user, I want to be able to trigger regeneration of specific SRD sections without affecting the entire document or disrupting system performance. Backend should expose an API and the frontend should provide a control to select and re-run individual SRD sections.

As a user, I want a Chat page where I can ask questions and receive concise, friendly answers in Hindi from the AI assistant. The page should display the avatar with emotional expressions (smiles, thinking, laughs), support step-by-step explanations, show occasional engaging questions from the assistant, and allow smooth back-and-forth dialogue. Implement based on the existing Chat (v2) JSX design. Receives input from Home page; leads to Image Prompt page on image request.

As a user, I want an Image Prompt page where I can request and view detailed, realistic, high-quality image prompt descriptions generated by the AI assistant. The page should clearly display the generated prompt text and a suggested visual description. Implement based on the existing Image Prompt (v2) JSX design. Navigated to from the Chat page when user requests an image.

As a user, I want the Chat page to be connected to the backend Chat API so that messages are sent and AI-generated Hindi responses are displayed in real time. Include loading states, error handling, and avatar emotional expression updates based on response content.

As a user, I want the Image Prompt page to be connected to the backend Image Prompt API so that generated image descriptions are fetched and displayed dynamically when I request an image from the Chat page.

As a user, I want the AI assistant's avatar to display dynamic emotional expressions—such as (smiles), (thinking), (laughs), (winks), (shrugs)—that update in real time based on the content and context of the assistant's responses, creating a more human-like and engaging interaction on the Chat and Home pages.