spring-openproject-mcp-server
MCP server to manage OpenProject work-packages realized in Java.
spring-openproject-mcp-server
MCP server to manage OpenProject work-packages realized in Java. The server acts as a proxy to your OpenProject API. The user's' OpenProject API token is used for authentication (not stored in the container!).
Open project compatibility
Tested against OpenProject 14,15,16.6,17.2
Get started (using LM Studio)
- Launch the Docker container
docker run -d -p 0.0.0.0:8080:8080 -e OPENPROJECT_URL=https://${$yourOpenProject} --tmpfs /tmp spring-openproject-mcp-server:latest - Run LM Studio and choose an appropriate model for your project domain. qwen/qwen3-coder-30b works well to create technical epics and user stories. For simple translations and text refinements, a smaller model is enough.
- Configure LM Studio's
mcp.jsonfile below "mcpServers":{...} and set the API token for the project to use (see below) - If you see "mcp/openproject-mcp" on the right side, below Integrations start prompting.
- Examples:
- Prompt: "List all projects in OpenProject using mcp/openproject-mcp."
- Prompt: "Use mcp/openproject-mcp to get work-packages from project id=10 and analyze the project."
- Prompt: "Use mcp/openproject-mcp and translate all epics in the project 'MyApp' to German."
- Prompt: "You are a precise Agile assistant. Help with the specification of the project. Use mcp/openproject-mcp and project id=10. Add INVEST criteria. Add Gherkin acceptance criteria. Add Definition of Done (checklist). Add Risks & Mitigations. Goals & measurable KPIs/OKRs (SMART, incl. target values)"
- Prompt: "Create 10 Epics for a new WebApp which does the following: ... Use mcp/openproject-mcp and project 'NewWebApp'."
Docker
Find the latest image at https://hub.docker.com/r/tmskln/spring-openproject-mcp-server
build
# convert java SBOM to optimized OCI format
docker run --rm \
-v "./target/classes/META-INF/sbom:/work" --platform linux/amd64 \
cyclonedx/cyclonedx-cli \
convert \
--input-file /work/application.cdx.json \
--output-file /work/application.spdx.json \
--output-format spdxjson
PKG_VERSION="dev" && docker build \
-f docker/Dockerfile \
-t spring-openproject-mcp-server:${PKG_VERSION} \
--attest type=sbom,generator=docker/scout-sbom-indexer:latest \
--label org.opencontainers.image.build-date=$(date -u +"%Y-%m-%dT%H:%M:%SZ") \
--label org.opencontainers.image.revision=$(git rev-parse HEAD) \
--label org.opencontainers.image.version=${PKG_VERSION} \
--platform linux/amd64,linux/arm64 \
.
run
docker run -d -p 0.0.0.0:8080:8080 -e OPENPROJECT_URL=https://${$yourOpenProject} --tmpfs /tmp spring-openproject-mcp-server
{
"mcpServers": {
"openproject-mcp": {
"url": "http://127.0.0.1:8080/sse",
"headers": {
"Authorization": "Bearer {YourOpenProjectApiToken}"
}
}
}
}
If you want to control the OpenProject server from the MCP-client or run against multiple OpenProject servers set start the container with -e OPENPROJECT_ALLOW_HEADER_BASE_URL=true and set the server URL in MCP config:
{
"Authorization": "Bearer {YourOpenProjectApiToken}",
"X-OpenProject-Base-Url": "https://${$yourOpenProjexct}"
}
Integration Tests
mvn test -Dopenproject.container.tag=16 -Dopenproject.container.port=18080
Mind to remove volumes between tests
TODO
- MCP OpenProject with OTEL
- use results after patch JSON
- add kubernetes deployment+service (helm?)
SSE vs. Streamable
spring.ai.mcp.server.protocol=SSE spring.ai.mcp.server.protocol=STREAMABLE
Spring creates one MCP server bean with one transport:
- SSE protocol → exposes:
- GET /sse for server → client streaming
- POST /mcp/message for client → server communication
- STREAMABLE protocol → exposes a single bidirectional endpoint (default /mcp) based on the new MCP streamable HTTP spec.
Because the transport defines the wiring, endpoints, and message routing, Spring cannot bind two protocols simultaneously.
FAQ
- Why wasn't a generated OpenAPI client used, but JSONNode value mapping with MapStruct?
- The first approach was using a generated OpenAPI client, but there were many issues:
- code generation with org.openapitools:openapi-generator-maven-plugin wasn't totally clean and needed lots of manual corrections.
- client is very strict and would fail minimal spec changes
- Spec lacks values, e.g. 'storyPoints' and manually enhancing generated code is not applicable
- compatibility to a wider range of versions can better be realized with direct value mapping and integration tests against multiple versions of OpenProject
- The first approach was using a generated OpenAPI client, but there were many issues:
License
This project is licensed under the MIT License – see the LICENSE file for details.
Related Servers
agentic-store-mcp
Power up your AI agents with 31 production-ready tools. Features local-first Python analysis, real-time SearXNG search, and a secure local proxy to sanitize prompts. Built for developers who value performance and privacy. Install once, use everywhere.
Summarize MCP
Converts text summaries to speech using OpenAI's Text-to-Speech API and plays them in the background.
Google Calendar Integration Project
Manage and interact with Google Calendar events using the Google Calendar API.
TaskFlow
An MCP server for integrating task management capabilities into Claude Desktop.
TaskMaster Todoist
An MCP server for integrating Todoist with Cursor AI. Requires a Todoist API token.
DeepWriter
Interact with the DeepWriter API, an AI-powered writing assistant.
Airfield Directory
Provides airfield aeronautical information for general aviation private pilots like runways, fuel prices, landing fees, weather, webcams and more
Backcast MCP Server
The Outcome Backcasting MCP is a strategic planning tool that helps you work backwards from a desired future outcome to identify the specific steps, resources, and dependencies needed to achieve your goals. Unlike traditional forward planning, backcasting starts with your end goal and creates a reverse roadmap to get there.
cwtwb
ai generate tableau file
SPAIK AI ROI
Predict and track AI ROI using Monte Carlo simulations, real-time industry benchmarks, and ML-powered insights.