yokai
Simple, modular, and observable Go framework for backend applications.
claude mcp add --transport stdio ankorstore-yokai docker run -i ankorstore/yokai
How to use
Yokai is a modular, observable Go framework designed to simplify building production-grade backend applications. This repository provides MCP (Model Context Protocol) support for Yokai by exposing its capabilities as a server that can be managed and orchestrated alongside other MCP-enabled services. The Yokai MCP server relies on preloaded core instrumentation (logging, tracing, metrics, health checks) and a dependency injection system, and it can be extended with additional modules to expose public HTTP/gRPC endpoints, workers, ORM integrations, and more. Once running, you can leverage Yokai’s module system to compose your application logic from built-in modules or your own custom modules, all while benefiting from standardized observability and configuration management.
To use this MCP server, start the container (or whichever deployment method you choose) and connect clients to the exposed endpoints as configured by Yokai and its modules. The server’s architecture allows you to wire in extension modules that provide specific capabilities (HTTP servers, gRPC services, background workers, database access, etc.) and rely on Yokai’s dependency injection to compose service behavior. You can also explore demonstrator or showroom projects within Yokai’s ecosystem to see concrete MCP-driven examples in action.
How to install
Prerequisites:
- Docker installed and running on your machine/server
- Optional: access to Yokai documentation for module selection and configuration
Installation steps:
- Pull or build the MCP container image:
- If using the published image: docker pull ankorstore/yokai
- If building locally (source code available): docker build -t ankorstore/yokai .
- Run the MCP server using Docker:
- docker run -d --name yokai-mcp -p 8080:8080 -p 4317:4317 ankorstore/yokai
- Verify the server is running by checking logs:
- docker logs -f yokai-mcp
- Optional configuration:
- If you have custom modules or configuration, mount configuration files or pass environment variables as needed when you start the container. See additional notes for common options.
Additional notes
Notes and tips:
- This MCP server uses Yokai's core instrumentation (logging, tracing, metrics) and a dependency injection system. Extend or replace modules to tailor functionality to your application needs.
- Common environment variables may include configuration paths, module selectors, and endpoints. Refer to Yokai documentation for exact variables supported by your modules.
- If you deploy in Kubernetes or another orchestrator, consider creating a Deployment/StatefulSet with the appropriate port mappings and health checks. Expose the necessary ports for the HTTP/GRPC endpoints provided by enabled modules.
- When debugging, check that the MCP server has the necessary permissions to access any external services (databases, message queues) and that network policies allow the required traffic.
- The MCP server can be extended with additional Yokai modules or contrib modules. Look for examples in the Yokai showroom to understand common module patterns and wiring.
Related MCP Servers
agentcontrolplane
ACP is the Agent Control Plane - a distributed agent scheduler optimized for simplicity, clarity, and control. It is designed for outer-loop agents that run without supervision, and make asynchronous tool calls like requesting human feedback on key operations. Full MCP support.
k8s
Manage Your Kubernetes Cluster with k8s mcp-server
sandbox
A Model Context Protocol (MCP) server that enables LLMs to run ANY code safely in isolated Docker containers.
github-brain
An experimental GitHub MCP server with local database.
mcp-tts
MCP Server for Text to Speech
tasker
An MCP server for Android's Tasker automation app.