generate_image
Generate an image using the Flux model
LABSCO SUMMARY
Image generation is one call, with the model pinned by configuration rather than chosen per prompt.
One tool and five arguments: prompt, aspect_ratio, num_outputs, output_format and seed. The model comes from the MODEL environment variable, so it is decided once at configuration time and every generation in the session uses it — predictable, and not something a prompt can change by accident. There is no editing, upscaling or history here: it generates and returns.
What it is
A one-tool server for image generation through Replicate, using the Flux model.
What you get
- An image generated from a text prompt, with the aspect ratio and output format chosen per call
- A seed argument, so a prompt and seed pair can be re-run and varied deliberately rather than by chance
- num_outputs, to take more than one candidate back from a single call
Requirements
REPLICATE_API_TOKEN and MODEL, and a Replicate account the token belongs to. MIT.
Setup effort
One command plus a key — npm install -g @gongrzhe/image-gen-server, then supply credentials
USE CASES
01
Getting an illustration for something you are writing
Describe the image, set the aspect ratio for where it will sit, and take back more than one candidate in the same call so there is something to choose between.
02
Iterating on a look you want to keep
Once a result is close, hold its seed and change one thing at a time — the prompt wording or the aspect ratio — so each attempt is a comparison rather than a fresh roll.
1 tool
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
BEFORE YOU INSTALL
What to know first
Tools you get
1 tool when we connected. The MODEL variable picks the Replicate model and defaults to black-forest-labs/flux-schnell; alternative-model-name in the example config is a placeholder. The tool returns image URLs, not image data.
Requires
- npx on PATH
- REPLICATE_API_TOKEN in your environment
Where it runs
Both documented setups run the server on your machine.
CONTEXT COST
Light Standing cost
Rated by the size of the one tool definition that rides along on every turn
INSTALLATION
Connect your client
Server setup
TERMINAL
npm install -g @gongrzhe/image-gen-serverMCP Clients
Claude Desktop — npx
CLAUDE DESKTOP — NPX · JSON
{
"mcpServers": {
"image-gen": {
"command": "npx",
"args": [
"@gongrzhe/image-gen-server"
],
"env": {
"REPLICATE_API_TOKEN": "<YOUR_REPLICATE_API_TOKEN>",
"MODEL": "alternative-model-name"
},
"disabled": false,
"autoApprove": []
}
}
}Claude Desktop — Local build
CLAUDE DESKTOP — LOCAL BUILD · JSON
{
"mcpServers": {
"image-gen": {
"command": "node",
"args": [
"<YOUR_PATH_TO_IMAGE_GEN_SERVER>/build/index.js"
],
"env": {
"REPLICATE_API_TOKEN": "<YOUR_REPLICATE_API_TOKEN>",
"MODEL": "alternative-model-name"
},
"disabled": false,
"autoApprove": []
}
}
}
