Overview
What is Bio-OS MCP Server?
Bio-OS MCP Server is a Model Context Protocol (MCP) based tool and prompt server for Bio-OS that provides workflow management and Docker image building capabilities. It is intended for users who need to interact with Bio-OS workflows and Docker image builds through an MCP-compatible client like CLINE or Claude Desktop.
How to use Bio-OS MCP Server?
Install dependencies (uv, Cromwell), clone the repository, and configure the server’s path and environment variables (MIRACLE_ACCESS_KEY, MIRACLE_SECRET_KEY) in CLINE’s MCP settings. Alternatively, use a pre‑configured Docker image in a Miracle Cloud IES instance. For optimal experience with CLINE, copy the contents of bioos-mcp-prompt.md into CLINE’s Custom Instructions.
Key features of Bio-OS MCP Server
- Submit and monitor Bio-OS workflows
- Upload and validate WDL workflow scripts
- Generate WDL input file templates
- Build Docker images and check build status
- Retrieve Docker build progress and logs
- Provides five MCP prompts for guided development
Use cases of Bio-OS MCP Server
- Submitting a Bio-OS workflow with an input JSON
- Validating a WDL script before submission
- Generating a template for WDL input parameters
- Building a Docker image from a Dockerfile or archive
- Monitoring the status of an ongoing Docker build
FAQ from Bio-OS MCP Server
What dependencies does Bio-OS MCP Server require?
It requires uv (Python package manager) and Cromwell (workflow execution engine).
How do I configure the server environment?
Set MIRACLE_ACCESS_KEY and MIRACLE_SECRET_KEY in the environment, and specify the absolute paths to uv, the server script, and the source directory in CLINE’s MCP settings.
Are MCP prompts supported in CLINE?
No, CLINE does not yet support MCP Prompts. For the best experience, copy bioos-mcp-prompt.md into CLINE’s Custom Instructions.
What deployment options are available?
You can install locally on your machine or use a pre‑configured Docker image inside a Miracle Cloud IES instance accessed via Code Server.
Which tools does the server expose?
It exposes six tools: submit_workflow, import_workflow, validate_wdl, generate_inputs, build_docker_image, and check_build_status.
