Vaadin

You can access and use an MCP server that enables IDE-assisted, context-aware search over Vaadin documentation. This server provides hierarchical navigation, framework-aware filtering, and standardized client interactions so your IDE can query, retrieve, and explore documentation efficiently.

How to use

You will connect your MCP client (IDE assistant) to the remote Vaadin MCP server to perform context-aware searches over Vaadin documentation. Start by configuring the client transport to reach the MCP endpoint, then perform searches that leverage hierarchical navigation to drill from specific details up to broader context. Use the available tools to retrieve full documents and to fetch versions or metadata as needed. The system will filter results by framework (Flow Java or Hilla React) and present context breadcrumbs to help you understand how results relate within the documentation structure. When exploring, prefer results that preserve parent-child relationships and show cross-file links for complete context.

How to install

Prerequisites you need before starting are the Bun runtime (version 1.3.6 or compatible) and API keys for the required services.

Install and set up the MCP server components using the provided workspace setup and run commands.

# Clone the project and install dependencies
git clone https://github.com/vaadin/vaadin-documentation-services
cd vaadin-documentation-services
bun install

# Build and prepare packages as needed (examples shown in standard flow)
bun run build

# Environment setup with required API keys
# Create a .env file with your API keys

Additional notes

The MCP server is available remotely via HTTP at the following endpoint. Use this URL to configure your MCP client’s transport to communicate with the server.

HTTP endpoint for MCP server: https://mcp.vaadin.com/

Security and maintenance

Use secure transport and rotate API keys regularly. Monitor health endpoints and set up automated alerts for processing or connectivity failures. Ensure you keep dependencies up to date and follow best practices for authentication and access control within your IDE integrations.

Troubleshooting

If you encounter connectivity or authentication issues, verify that the MCP endpoint URL is reachable from your network and that your API keys are correctly configured in the environment file. Check the health endpoint for status and review logs for any error messages related to embedding or document retrieval.

Examples and capabilities

The MCP server exposes tools to search Vaadin docs and to fetch full documents. You can query with semantic and keyword matching, retrieve a complete page, and obtain the latest Vaadin version metadata. The framework-aware filtering ensures you see only content relevant to Flow (Java) or Hilla (React) as needed.

Environment and tooling

Environment variables used by the server include the keys required for embeddings and vector storage. You supply them in your environment configuration so the MCP server can access the embedding service and vector database. These variables must be kept secret and not hard-coded into client configurations.

Available tools

search_vaadin_docs

Search Vaadin documentation using semantic and keyword matching to return relevant results with framework-aware filtering.

get_full_document

Retrieve the complete documentation page with all sections and context.

get_vaadin_version

Fetch the latest Vaadin version and its release timestamp.