Duply
← Image automation guides

Agent workflow

Create Images with MCP

Learn how an MCP-compatible agent can call a small tool that maps prompts or structured data to Duply image templates.

Quick answer

Create images with MCP by exposing a narrow tool that accepts a template, validates dynamic values, calls Duply's image API, and returns the generated image URL to the agent.

Best-fit use cases

AI assistants and agent workflows
Internal content agents
Prompt-to-template generation
Human-reviewed automation

Implementation

A reliable four-step workflow

1

Define the tool

Use a clear tool name such as generate_branded_image with a strict input schema.

2

Map templates

Let the agent select only from approved template identifiers and documented layer names.

3

Call Duply

The MCP tool validates inputs and sends a normal request to the Duply image API.

4

Return evidence

Send the final image URL, template used, and input summary back to the agent for review.

Concrete example
trigger
An agent prepares a campaign brief
inputs
Template ID, headline, photo URL, CTA
output
A deterministic branded image
destination
Approval step or campaign workflow

Frequently asked questions

Does MCP replace the image API?

No. MCP is the agent-facing tool layer; the underlying generation request can still use Duply's image API.

Should an agent be allowed to edit every layer?

Usually no. Expose only approved templates and dynamic fields so the agent cannot change locked brand elements.

Continue the workflow

Create images with an API →Dynamic image URLs →All image automation guides →

Build one production-shaped example

Start with one approved template, realistic data, and the exact trigger your team intends to use.

Start free