figurify-mcp 0.2.2

Ship-ready MCP server for Gemini + Imagen diagram/plot generation. Runs on Claude, ChatGPT, VSCode Copilot, Cursor.

figurify-mcp

An MCP server that adds image generation to Claude, Cursor, VSCode Copilot, and any other MCP-compatible host that does not have native image generation.

Currently supports Google Gemini + Imagen. Support for other image-generation providers is on the roadmap.

Install

pip install figurify-mcp

You will need a Gemini API key — create one at aistudio.google.com/apikey.

Setup in your host

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "figurify": {
      "command": "uvx",
      "args": ["figurify-mcp"],
      "env": { "GOOGLE_API_KEY": "your-key" }
    }
  }
}

Claude Code / Cursor

claude mcp add figurify -s user \
  -e GOOGLE_API_KEY=$GOOGLE_API_KEY \
  -- uvx figurify-mcp

VSCode Copilot

Create .vscode/mcp.json:

{
  "servers": {
    "figurify": {
      "command": "uvx",
      "args": ["figurify-mcp"],
      "env": { "GOOGLE_API_KEY": "your-key" }
    }
  }
}

Restart the host. Ten tools become available automatically.

Tools

Tool	What it does
create_simple_diagram	Prompt → image. One call.
create_infographic	Topic → styled poster (VLM designs layout, Imagen renders).
create_methodology_diagram	Academic paper figure with iterative critique.
create_statistical_plot	Tabular data → matplotlib chart.
evaluate_diagram	VLM-as-judge comparison of two diagrams.

Five additional meta-tools help the host pick the right tool without loading every schema upfront: search_tools, describe_tool, recommend_model, estimate_cost, list_available_models.

All generation tools default to 1K resolution so images render inline without hitting the ~1 MB tool-result cap enforced by Claude.

Configuration

All environment variables are optional except GOOGLE_API_KEY.

Variable	Default	Description
GOOGLE_API_KEY	required	Gemini + Imagen API key
FIGURIFY_OUTPUT_DIR	~/.figurify/outputs	Where generated PNGs are saved
FIGURIFY_VLM_MODEL	auto-routed	Pin a specific Gemini text/vision model
FIGURIFY_IMAGE_MODEL	auto-routed	Pin a specific image-generation model
FIGURIFY_MAX_METHODOLOGY_ITERATIONS	4	Upper bound for the critique loop

Transports

figurify-mcp                                             # stdio (default)
figurify-mcp --transport http --host 0.0.0.0 --port 8765 # streamable HTTP

Stdio is what hosts like Claude Desktop use. HTTP is useful for remote deployments (ChatGPT Apps SDK, VSCode Copilot remote MCP).

Development

git clone https://github.com/genaimanoj/figurify-mcp
cd figurify-mcp
uv venv && source .venv/bin/activate
uv pip install -e '.[dev]'
cp .env.example .env   # set GOOGLE_API_KEY
pytest -q

Issues, feature requests, and pull requests welcome at github.com/genaimanoj/figurify-mcp.

License

MIT · © 2026 Manoj Bhandari