AI coding agents

A coding agent such as Claude Code, Codex or Cursor can call the Garmo API for you: it finds your product photos, sends them, waits for the job and saves the finished images in a folder. You describe what you want; you do not write the code.

1. Give the agent a key

Put an API key in the environment the agent runs in, never in a prompt or a file you commit. Start with a free test key: it runs everything and returns your own photo as the result.

Shell
export GARMO_API_KEY=gm_test_…

2. Point it at the docs

Every docs page is also plain Markdown, at the same address with .md (for example /api/docs/generate.md), and the whole documentation is one file at /api/docs/llms-full.txt. A prompt to start from:

Prompt
Read https://garmo.studio/api/docs/generate.md, then make two on-model photos of ./dress.jpg and save them in ./photos

3. Or install the Garmo skill

The skill is one SKILL.md file that teaches an agent the whole workflow: upload, create, wait, download. Claude Code reads skills from ~/.claude/skills/, Codex from ~/.codex/skills/, and Cursor reads both. This puts it in the two folders:

Shell
for d in ~/.claude/skills ~/.codex/skills; do
  mkdir -p "$d/garmo" && curl -fsSL https://garmo.studio/api/skill/SKILL.md -o "$d/garmo/SKILL.md"
done

Then ask in plain words, in any language: "make two on-model photos of every photo in ./new-arrivals". The agent uses the skill when the task matches it. The file: /api/skill/SKILL.md.

Other agents

Gemini CLI, GitHub Copilot and any agent that can make HTTP requests can use the Markdown docs, the OpenAPI 3.1 file or the skill the same way.

An agent spends credits like any other client. Build with a test key, then switch to a live key; a live key's limits and balance are on GET /account.