A playful art lab for transforming your likeness into authentic, museum-style artwork
MuseoMorph is an open-source desktop app that transforms your photo into artwork inspired by the great artists and movements of history.
This project grew out of months of experimenting with Google’s NanoBanana image model. Early attempts with loose prompts (“make me look like Van Gogh”) produced outputs that felt plasticky and modern — closer to filters than paintings. Through trial and error, I learned that:
- Loose prompts → generic defaults.
- Tight, detailed prompts → better authenticity.
- Iterative refinement is essential.
MuseoMorph encodes those lessons into the Universal Museum-Grade Prompt Framework (v4.2) — a scaffold designed to respect artists’ materials, methods, and contexts.
MuseoMorph is not about counterfeiting or reproducing real works. No frames, no logos, no signatures. It’s about playful interpretation — a toy-like art lab that also invites users to learn about art history.
- Museum-inspired UX: Landing page with collections, gallery-style style pages, and clean detail views.
- Markdown-driven styles: Each
.mdfile encodes ratios, modes, flags, and educational blurbs. Drop one in → app auto-detects it. - Multi-subject support: Works with single portraits or group photos (manual face detection by default; auto toggle available).
- Respectful prompts: Guardrails for age, gender, wardrobe, and IP safety.
- Local-first: Runs fully on your computer. No image uploads.
- Your API key, your control: Start with Gemini/NanoBanana. Other providers may be added later.
Each collection and style page includes short educational blurbs:
- When the movement began
- Key figures and themes
- What made it distinctive
Click into an artist for more detail. Think of it as browsing the Met or MoMA websites — curated and inviting — with the bonus that you can put yourself in the art.
MuseoMorph is a playful art lab: fun first, but also a way to appreciate the craft of artists.
- Download MuseoMorph (cross-platform builds coming via Tauri).
- Add your API key (Gemini/NanoBanana to start).
- Upload a photo → Pick a style → Generate.
index.html– Vite entry point served during dev/build.public/– static assets copied verbatim into output.src/– React + Vite UI shell that surfaces collections from the generated manifest.src-tauri/– Rust backend bootstrapped for the Tauri desktop wrapper.scripts/– Node/TypeScript utilities for manifest generation and validation.styles/– curated Markdown prompt frameworks consumed by the app.styles-source/- long-form research notes and drafting material for prompts.
- Keep one artist or one tightly scoped style per file.
- Use
style_scopecorrectly:single-artist,movement, orcollaboration.- Duos/studios (e.g., “Mouse & Kelley”, “Hipgnosis”, “Siegel & Shuster”) → set
style_scope: "collaboration"(do not split). - Schools/workshops (e.g., “Utagawa School”, “Morris & Co. Workshop”) →
style_scope: "movement".
- Duos/studios (e.g., “Mouse & Kelley”, “Hipgnosis”, “Siegel & Shuster”) → set
- Ratios: ≤3, follow docs/ratio_policy.md; start as
provisionalwith a brief rationale. - Safety: no nudity/gore/logos/watermarks; minors modest; gender respect.