generate_imageLABSCO SUMMARY
DALL-E with a file path, which is the part that matters.
Images land as files where you tell them to, so the next step in the conversation can pick them up instead of looking at a URL. Two capabilities are DALL-E 2 only — editing an image and making variations of one — so a workflow that starts on DALL-E 3 cannot finish there.
What it is
A Node server over OpenAI's DALL-E API. generate_image takes a prompt and writes the result to disk; edit_image works on an image that already exists.
What you get
- generate_image with model, size, quality, style and count as parameters — DALL-E 3 or DALL-E 2
- Sizes that follow the model: 1024x1024, 1792x1024 or 1024x1792 on DALL-E 3; smaller squares on DALL-E 2
- edit_image to modify an existing image, DALL-E 2 only
- Variations of an existing image, DALL-E 2 only
- Output location set per call with saveDir and fileName, or globally in the environment
- A check that the OpenAI API key is valid
Requirements
Node, and a clone of the repository built with npm. An OpenAI API key in the environment, alongside the default output folder. In Cline, set saveDir to the current workspace directory so the editor can find and display what was generated.
Setup effort
Build from source — git clone https://github.com/Garoth/dalle-mcp.git
USE CASES
01
Getting an illustration into the project you are working in
Generate with saveDir pointed at the workspace and a fileName you choose, so the file is where the rest of the work is.
02
Trying the same idea at two aspect ratios
Run the prompt at 1024x1024 and again at 1792x1024 and compare — the sizes available depend on which model you asked for.
03
Changing part of an image you already have
Edit the existing file rather than re-prompting from scratch; this path runs on DALL-E 2.
04
Checking the key works before a batch
Call validate_key first, so a bad OpenAI API key surfaces before ten generations fail.
4 tools
Listed in project docs. Your client may show a different list.
edit_imagecreate_variationvalidate_keyBEFORE YOU INSTALL
What to know first
Tools you get
4 tools named in the project's documentation.
Requires
OPENAI_API_KEY in your environment
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
# Clone the repository
git clone https://github.com/Garoth/dalle-mcp.git
cd dalle-mcp
# Install dependencies
npm install
# Build the project
npm run buildMCP Clients
Cline
CLINE · JSON
{
"mcpServers": {
"dalle-mcp": {
"command": "node",
"args": [
"<YOUR_PATH_TO_DALLE_MCP>/build/index.js"
],
"env": {
"OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>",
"SAVE_DIR": "<YOUR_SAVE_DIR>"
},
"disabled": false,
"autoApprove": []
}
}
}<YOUR_PATH_TO_DALLE_MCP> is the full path of the folder you cloned, and <YOUR_SAVE_DIR> is the folder images should be saved to.
