Can I vibecode Flair AI?

Build a highly polished, local-first generative media workspace tailored for product photography, acting as a personal replacement for Flair AI. Vibe-code this application using Python 3.12, FastAPI, and SQLite for the backend, integrating with ComfyUI as a local rendering worker. The frontend must be a beautiful, highly responsive React application (Vite + Tailwind CSS). The UI should look like a cutting-edge creative tool: dark mode by default, sophisticated control panels, smooth image loading transitions, and a premium, professional feel.

Core Workflow & Features:
1. **Scene Composition:** Build a drag-and-drop canvas where users can compose product scenes. Users should be able to upload product cutouts, arrange digital props, and define areas for generated backgrounds.
2. **Local Generation Engine:** The backend must submit generation jobs exclusively to a user-owned local ComfyUI endpoint (configured via `.env`). Provide UI controls for prompts, negative prompts, seeds, dimensions, models, and specific ComfyUI workflows.
3. **Reproducibility & Provenance:** For every output generated, record the exact generation parameters and the full ComfyUI workflow JSON in the local SQLite database. Ensure strict reproducibility.
4. **Asset Management:** Build a highly visual, searchable contact sheet for generated images. Include features to compare, favorite, annotate, and quickly re-run variations of previous generations.
5. **Local-First Privacy:** All user data, uploads, and generated images must remain strictly on the local filesystem. Do not add accounts, billing, telemetry, analytics, or a hosted control plane. Support local image-to-image and mask inputs securely.
6. **Robust UX:** Show estimated VRAM requirements and fail gracefully with clear error messages when a workflow or model is missing. Include beautiful empty states, loading skeletons, and success validations. 
7. **Engineering Standards:** Make the first run work locally with one documented command. Write focused tests for the backend API and one end-to-end happy path. Create a comprehensive README covering setup, ComfyUI integration, permissions, and data backup. Deliberately exclude public generation hosting, model training, and copying proprietary vendor datasets.