Install kie-models

Let your coding agent build with 200+ image, video, music and speech models on KIE.

kie-models teaches your agent how to use the KIE model market: find a model, read its parameters, upload input files, submit the task, poll until it finishes and save the result.

Prerequisites

A coding agent that supports skills, such as Claude Code, Codex or Grok Build.

Node.js, so that npx is available.

A KIE API key. Create one at kie.ai/api-key.

Step 1: Install the skills

Run once per machine:

This installs both kie-models and kie-chat-agents. Run the same command again later to update to the latest version.

Step 2: Set your API key

The skill reads your key from the KIE_API_KEY environment variable. It never needs the key written into your project.

macOS and Linux

Export it in the shell that starts your agent, or add it to your shell profile (~/.zshrc, ~/.bashrc):

Windows (PowerShell)

$env:KIE_API_KEY = "your_key"
setx KIE_API_KEY "your_key"

setx only applies to windows opened afterwards. Set $env: as well if you want to use it right away.

Step 3: Try it

Restart your agent from a shell where KIE_API_KEY is set, then ask in plain words:

Use the kie-models skill to run gpt-image-2-text-to-image on KIE. Prompt: "a white ceramic mug on a light oak table, soft daylight, product photo". Aspect ratio 1:1, resolution 1K. Save the image to ./output.

The agent searches the live model list, reads the model's parameters, submits the task, waits for it and saves the file to ./output.

Security

Keep your API key in an environment variable. Do not paste it into chat, screenshots or files that go into a repository.

If your key leaks, reset it from kie.ai/api-key.

Prerequisites#

Step 1: Install the skills#

Step 2: Set your API key#

macOS and Linux#

Windows (PowerShell)#

Step 3: Try it#