Get the FREE Ultimate OpenClaw Setup Guide →

kroki

Kroki-MCP is a Go-based Model Context Protocol tool that converts textual diagram definitions (PlantUML, Mermaid, and more) into images via a Kroki backend. Designed for simplicity and flexibility, it supports both local and remote Kroki servers, offers configurable settings, and outputs multiple formats – making it ideal for developers building AI

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio utain-kroki-mcp go run github.com/utain/kroki-mcp/cmd/kroki-mcp@latest -m stdio -f png --kroki-host https://kroki.io

How to use

Kroki-MCP exposes a command-line MCP tool that uses Kroki to render textual diagrams (such as PlantUML or Mermaid) into images. It can operate in two modes: SSE (default, streaming results) and STDIO (reads diagram code from stdin and writes the image to stdout). The MCP integration allows you to run Kroki-MCP as an MCP server and wire it into your MCP workflow, so clients can request diagram rendering via the MCP protocol. By default, Kroki-MCP targets a Kroki backend (default https://kroki.io) and outputs PNG, but you can change the output format to SVG, JPEG, or PDF and point Kroki-MCP at a custom Kroki server. To use it, run the Kroki-MCP binary with the desired options, or configure it in your MCP config as shown in the README examples. The MCP integration leverages the mcp-go library, enabling seamless tool/server integration within your MCP ecosystem.

How to install

Prerequisites:

  • Go 1.22 or newer installed on your system
  • Internet access to fetch modules
  • Optional: Docker if you want to run via Docker (as described in the README)

Install steps:

  1. Ensure Go is installed and in your PATH. Verify with: go version

  2. Run Kroki-MCP directly from source without building a separate binary (quick start): go run github.com/utain/kroki-mcp/cmd/kroki-mcp@latest -m stdio -f png --kroki-host https://kroki.io

  3. Build a local binary for repeated use: go build -o kroki-mcp ./cmd/kroki-mcp ./kroki-mcp -m stdio -f png --kroki-host https://kroki.io

  4. Running as an MCP server configuration (as shown in the README):

    • Create an MCP config file with the kroki-mcp server (see example in README)
    • Start the MCP server manager; Kroki-MCP will be invoked via Go as described above

Docker-based option:

  • Build the Docker image and run (as per the Development with Docker section): docker build -t kroki-mcp . docker run --rm -it kroki-mcp --help

Optional: refer to the project README for additional Docker Compose usage to connect to a Kroki server running locally.

Additional notes

Notes and tips:

  • The default Kroki backend is https://kroki.io; you can override this with --kroki-host or --kroki-host in the MCP config.
  • Output formats include png, svg, jpeg, and pdf; specify with -f/--format.
  • In STDIO mode, Kroki-MCP reads diagram code from stdin and writes the rendered image to stdout, which makes it easy to pipe data within a larger toolchain.
  • In SSE mode (the default in some setups), results stream as events, suitable for real-time processing.
  • When integrating with MCP, ensure the command and arguments align with your MCP server runner expectations (the example uses go run with @latest to keep it current).
  • If you modify flags or host settings, ensure your MCP client requests match the server’s accepted flags and formats.
  • Check connectivity to Kroki server backends if you’re self-hosting Kroki behind a firewall or proxy.

Related MCP Servers

Sponsor this space

Reach thousands of developers