coroot-mcp exposes Coroot incidents and application health summaries as MCP tools that an AI assistant can call via JSON-RPC over HTTP at /mcp. Tools available include list_recent_incidents for surfacing recent incidents with context, get_incident_details for detailed incident context, summarize_incident_root_cause for natural-language RCA explanations, and investigate_incident which returns an RCA summary along with structured data. You can also obtain an overview of project health with get_applications_overview and per-service health via get_service_health. This makes it possible to triage incidents, generate executive-friendly summaries, and enrich runbooks or dashboards with machine-readable Coroot data. To use, point your MCP-aware client to the /mcp endpoint (e.g., http://localhost:8080/mcp) and call the available tools, optionally filtering by project, service, or time range as supported by the server.

Prerequisites:

• JDK 21+ (for local development) and Maven Wrapper (or Maven) installed if you want to build from source.
• Java 21+ and a compatible environment if you intend to run the built JAR directly.
• Docker (for the provided docker-based run command) if you choose to deploy via Docker.

Install from source (recommended for contributing or local testing):

1. Clone the repository git clone https://github.com/Meritocra/coroot-mcp.git cd coroot-mcp

2. Build with Maven Wrapper (preferred) or Maven ./mvnw clean package -DskipTests

   or if you have mvn installed

   mvn clean package -DskipTests

3. Run the Spring Boot application locally (example using the built JAR): java -jar target/coroot-mcp-<version>.jar

   By default, the MCP endpoint will be available at http://localhost:8080/mcp

4. Optional: run with a stub Coroot profile for testing without a real Coroot instance: ./mvnw spring-boot:run -Dspring-boot.run.profiles=stub-coroot

Deploy with Docker (alternative):

1. Pull and run the image docker run -p 8080:8080 -e COROOT_API_URL=https://coroot.your-company.com -e COROOT_API_KEY=... meritocra/coroot-mcp:latest

2. Confirm the service is up by hitting health: curl -s http://localhost:8080/actuator/health curl -s http://localhost:8080/mcp -H 'Content-Type: application/json' -d '{"jsonrpc":"2.0","id":"init-1","method":"initialize","params":{}}'

Environment variables are used for connecting to Coroot and OpenAI. Ensure COROOT_API_URL, COROOT_API_KEY, and OPENAI_API_KEY are securely managed. The default port is 8080; adjust if needed. If you are using Docker, remember to expose port 8080 to your host. The OPENAI_MODEL value supports specific commands for RCA and natural-language summaries; choose a model your environment supports. Review mcp.json or in-repo documentation for additional tooling options and example configurations. If you encounter connectivity issues to Coroot, verify COROOT_DEFAULT_PROJECT_ID and network access from the MCP container or host.