Image Generation MCP Server

pvliesdonk/image-generation-mcp

Connects Claude to OpenAI DALL-E, Google Gemini image models, and local Stable Diffusion (via SD WebUI) through a single interface. Exposes `generate_image` with auto-routing based on prompt keywords, plus on-demand image transforms (resize, crop, format conversion) through `image://` URIs. Background task support for long SD generations, persistent event store for session resumability over HTTP, and a built-in gallery viewer. Ships with Docker, systemd packages, and .mcpb bundles for Claude Desktop. Reach for this when you want Claude to generate production images without writing custom API clients for each provider, or when you need to test prompts across multiple models from one conversation.

GitHub stars: 1
tool: 1
auth: Secret needed
transport: STDIO, HTTP
env vars: 24
last updated: Aug 25, 2026

View on GitHub

Tools

1 tools

Public tool metadata for what this MCP can expose to an agent.

image-generationGenerate an image from a given text10 params

Generate an image from a given text

Parameters* required

advance_configobject

Advanced configuration options.

aspect_ratiostring

The aspect ratio for the image.one of 1:1 · 16:9 · 21:9 · 3:2 · 2:3 · 4:5

file_store_keystring

Optional file store key.

heightnumber

The height of the image in pixels.

output_formatstring

The output format of the image.one of png · svg

promptstring

The prompt to generate the image.

return_typestring

The return type for the image.one of url · binary · base64

stepsnumber

The number of steps for image generation.

urlstring

Optional URL for image input.

widthnumber

The width of the image in pixels.

Configuration

IMAGE_GENERATION_MCP_KV_STORE_URLdefault: file:///data/state

Persistent-state backend URL shared by every pvl-core subsystem that needs state. `memory://` is in-process and lost on restart; `file:///path` persists on one server; `redis://`, `dynamodb://` and `mongodb://` each need their matching extra. When unset, defaults to `file:///data

IMAGE_GENERATION_MCP_TOOLS_ALLOW

Comma-separated explicit tool names this instance exposes; every other tool is hidden from listings and cannot be invoked. Names matching no registered tool are inert. Mutually exclusive with `tools_deny`. Takes effect through `apply_tool_visibility`.

IMAGE_GENERATION_MCP_TOOLS_DENY

Comma-separated explicit tool names hidden from this instance (absent from listings, cannot be invoked). Names matching no registered tool are inert. Mutually exclusive with `tools_allow`. Takes effect through `apply_tool_visibility`.

IMAGE_GENERATION_MCP_SERVER_NAME

Rename this server instance; defaults to the project name.

IMAGE_GENERATION_MCP_INSTRUCTIONS

Replaces the default MCP instructions text sent to clients.

FASTMCP_LOG_LEVELdefault: INFO

Log level for FastMCP internals and app loggers (DEBUG / INFO / WARNING / ERROR / CRITICAL). The -v CLI flag overrides to DEBUG.

FASTMCP_ENABLE_RICH_LOGGINGdefault: true

Set false for plain or structured JSON log output.

IMAGE_GENERATION_MCP_A1111_HOST

Deprecated alias for IMAGE_GENERATION_MCP_SD_WEBUI_HOST; logs a warning when used.

IMAGE_GENERATION_MCP_A1111_MODEL

Deprecated alias for IMAGE_GENERATION_MCP_SD_WEBUI_MODEL; logs a warning when used.

IMAGE_GENERATION_MCP_READ_ONLYdefault: true

When true, write-tagged tools (image generation, transforms, uploads) are hidden from clients. Set false to enable them.

IMAGE_GENERATION_MCP_SCRATCH_DIRdefault: ~/.image-generation-mcp/images

Directory where generated images are saved. Created automatically on first use.

IMAGE_GENERATION_MCP_OPENAI_API_KEYsecret

OpenAI API key. Enables the OpenAI provider (gpt-image-2, gpt-image-1.5, dall-e-3) when set.

IMAGE_GENERATION_MCP_GOOGLE_API_KEYsecret

Google API key. Enables the Gemini provider (gemini-3.1-flash-image and others) when set. Get a key at https://aistudio.google.com/apikey.

IMAGE_GENERATION_MCP_SD_WEBUI_HOST

SD WebUI base URL (such as http://localhost:7860). Enables the SD WebUI provider when set. Compatible with AUTOMATIC1111, Forge, reForge, and Forge-neo.

IMAGE_GENERATION_MCP_SD_WEBUI_MODEL

SD WebUI checkpoint name, used for model-aware preset detection (SD 1.5 / SDXL / Lightning) and checkpoint override. Unset uses the instance's current model.

IMAGE_GENERATION_MCP_DEFAULT_PROVIDERdefault: auto

Provider used when no keyword triggers auto-selection: auto, openai, gemini, sd_webui, or placeholder. auto picks the first configured provider.

IMAGE_GENERATION_MCP_TRANSFORM_CACHE_SIZEdefault: 64

Maximum number of transformed image results (resize, crop, convert) kept in memory. Set 0 to disable caching.

IMAGE_GENERATION_MCP_PAID_PROVIDERSdefault: openai

Comma-separated provider names that cost money; generate_image asks for confirmation (client elicitation) before using them. An empty value falls back to this default; to disable confirmation, set a value that names no provider (such as none).

IMAGE_GENERATION_MCP_STYLES_DIRdefault: ~/.image-generation-mcp/styles

Directory for style preset files (Markdown with YAML front matter). Created automatically if it does not exist.

IMAGE_GENERATION_MCP_ALLOW_LOCAL_FILE_INPUTdefault: false

Allow reading input images from local filesystem paths. Off by default: only URLs and uploads are accepted.

IMAGE_GENERATION_MCP_MAX_INPUT_IMAGE_BYTESdefault: 20971520

Maximum accepted input image size in bytes.

IMAGE_GENERATION_MCP_FETCH_TIMEOUT_Sdefault: 30.0

HTTP timeout in seconds when fetching remote image URLs (fetch_image and URL inputs).

IMAGE_GENERATION_MCP_TRANSFER_TTL_DEFAULT_Sdefault: 3600.0

Link lifetime in seconds when the caller requests no explicit TTL.

IMAGE_GENERATION_MCP_TRANSFER_TTL_MAX_Sdefault: 86400.0

Ceiling in seconds a caller-requested link TTL is clamped to.

Readme