Luma Photo Studio — Development skill for Claude Code
Drop one face reference, walk out with a 12-image photoshoot.
How to install Luma Photo Studio
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open AKCodez/luma-photo-studio and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Luma Photo Studio does
Drop one face reference, walk out with a 12-image photoshoot. Powered by Luma Uni-1 + Claude.
Alternatives in Development
- HTML Ppt Skill — HTML PPT Studio — AgentSkill with 24 themes, 31 layouts, 20+ animations for building professional HTML present 8.1k ★
- Lumenx Build — LumenX Studio 桌面应用构建流程 - macOS DMG 和 Windows EXE 打包 1.2k ★
- Claude Prism — An offline-first scientific writing workspace powered by Claude 1.1k ★
README
Luma Photo Studio
Drop one face. Walk out with a 12-image photoshoot.
**Live demo:** https://luma-photo-studio.vercel.app
A single-page studio that takes up to three reference photos of your face and produces a curated 12-image shoot. Three modes: pre-built **Packs** (GTA 6, Editorial, Dating Profile, Cinematic, Tech Founder, Streetwear, Travel, Dark Academia, Y2K, Album Cover, Magazine Cover, Anime Con, Simpsons-inspired), **Surprise Me** (Claude invents 12 distinct scenes from a one-word vibe), and **Custom** (write your own). Click any image to edit it in plain English.
Built on the [Luma Agents API](https://lumalabs.ai/uni-1) (`uni-1` model) for image generation + editing, Claude Sonnet for scene authoring, Next.js, and Vercel.
What it does
- Multi-reference upload — drop up to 3 photos for a stronger likeness lock
- 11 curated Packs — each pack auto-fills 12 hand-written scene prompts in a distinct aesthetic
- Surprise Me — type a one-word vibe ("hacker", "tokyo at 3am") and Claude composes 12 unique scenes
- Custom — write your own 12 prompts
- Plain-English editing — click any image, type "make the jacket red" and only that changes
- History — every past shoot is browsable at
/history - Permalinks —
/shoots/{id}for any shoot, shareable
Tech stack
- Next.js 15.5 App Router + React 19 + TypeScript strict
- Tailwind CSS + shadcn primitives + Framer Motion
- @anthropic-ai/sdk — Claude Sonnet 4.5 for surprise scene authoring (tool calling)
- Luma Agents API (
https://agents.lumalabs.ai/v1) —uni-1model withimage_reffor face lock andtype: "image_edit"+sourcefor edits - Vercel Blob — direct browser-to-blob uploads (bypasses 4.5 MB serverless body cap), per-image state stored as separate JSON blobs to avoid concurrent-write races
- React Query + Sonner for client polling and toasts
Quick start (local)
# 1. Clone
git clone https://github.com/AKCodez/luma-phoRelated Skills
Peeka
Agent-native studio for procedural 2D avatars. Humans shape a face in the browser; agents build one over the A
Technique Combine People
Combine a user's photo with another person (real or fictional) into a single composite image: $ARGUMENTS
Photo Clear
Clear a photograph for the book — the go/no-go gate on image rights, licensing, and caption juxtaposition. Own
Mads
The Multi-Agent Development Studio (MADS) is a vibe-coded programming environment for macOS which allows users
Conformance Walk
Conformance Walk — auto-detect the project flavor, run the matching CW adapter family (kailash reference engin
ㅇㅇ
Drop the current OS clipboard (text or image) into context and act on it — Hangul-mode alias of /dd (typing "d
Related Agents
Nancy Legal Risk Counsel
Use when reviewing passages involving living people, companies, active litigation, allegations, legal status,
Image Sourcer
Finds REAL, existing images on the web — stock photos, illustrations, icons, logos, textures — with correct li
Daily Copilot
The everyday work partner. Use for prioritization, thinking out loud, "remember this", "challenge this", "what
