MCP Imagenate — Claude Code MCP server: image generation from chat
Wires image and short-video generation into your coding agent through Google Gemini, Anthropic, BFL FLUX or Reve, with one install line and one set of keys.
claude mcp add mcp-imagenate -- npx -y mcp-imagenateWhat MCP Imagenate does
The server exposes two tools. generate_image takes a prompt plus a model name; generate_video makes 3–10 second clips with audio using Google Gemini Omni. Pick the provider by model: nano-banana-pro when quality matters most, gpt-image-2.5-flare as the default, flux-2-max from BFL, reve-image for typography-heavy work.
You need Node.js 22+ and a key for at least one provider. Keys come from environment variables — GEMINI_API_KEY or NANO_BANANA_API_KEY, OPENAI_API_KEY or GPT_IMAGE_API_KEY, BFL_API_KEY, REVE_API_KEY or REVE_API_TOKEN. Setting NANO_BANANA_OUTPUT_DIR is optional; it confines every read and write path to a single directory, which is the safer setup on a production box.
Reve behaves differently. Size is ignored, prompts above 4000 characters are refused, and at most eight reference images go through. A job takes 40–80 seconds, so any proxy or queue in front needs a timeout of at least 120 seconds. Transparent backgrounds come only from the OpenAI models.
What it helps with
- produce an illustration from a written prompt without leaving the chat
- render a short clip with audio for a draft deck
- switch image models without editing the config
- keep generated files inside one directory with
NANO_BANANA_OUTPUT_DIR - get a transparent-background asset for layout work
Example prompts for the agent
- Generate a header image for a post about coffee using
nano-banana-pro - Make a 5-second clip of waves at sunset with
generate_video - Draw a transparent icon with
gpt-image-2.5-sunburstand save it toassets/
How to install MCP Imagenate
Platform CLI
claude mcp add mcp-imagenate -- npx -y mcp-imagenatePlatform CLI
codex mcp add mcp-imagenate -- npx -y mcp-imagenateConfig file
{
"mcpServers": {
"mcp-imagenate": {
"command": "npx",
"args": [
"-y",
"mcp-imagenate"
]
}
}
}Config file
{
"servers": {
"mcp-imagenate": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"mcp-imagenate"
]
}
}
}Platform CLI
gemini mcp add mcp-imagenate npx -y mcp-imagenateStep by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about MCP Imagenate
How do I install MCP Imagenate in Claude Code?
Run claude mcp add mcp-imagenate -- npx -y mcp-imagenate. Add at least one provider key to the environment, or the server has nothing to call.
What does MCP Imagenate do?
It gives the agent generate_image for stills and generate_video for 3–10 second clips. Requests go to Google Gemini, OpenAI, BFL FLUX or Reve depending on the model you name.
Which models can I choose?
Gemini offers nano-banana-2, nano-banana-2.1 and nano-banana-pro; Anthropic offers gpt-image-2.5-flare, gpt-image-2.5-sunburst and gpt-image-2; BFL FLUX offers flux-2-klein, flux-2-pro and flux-2-max, and reve-image covers the Reve provider.
