MCP-scala

MCP-scala is an MCP server and framework implementation written in Scala 3 that allows developers to create and serve Model Context Protocol capabilities on top of the Scala and JVM ecosystem. It connects MCP-compliant clients to custom tools and logic defined using functional Scala libraries such as Cats Effect, Circe, and Tapir. Designed for JVM and Scala developers, it enables the integration of typed backend services into LLM-driven environments. Users can define tool schemas automatically derived from case classes and expose functions using standard input/output transport. The project is currently in an alpha development stage, offering automatic JSON Schema derivation, tool definition mechanisms, text content handling, and stdio transport. Example implementations provided in the repository include basic arithmetic and sequence operations like randomNumber, iota, and sum, demonstrating how developers can wrap computational routines and expose them directly to AI agents like Cline.

Category: AI & LLM Tooling

Tags: framework, jvm, scala

Visit MCP-scala

How to install and configure MCP-scala

1. Build the server using sbt to compile the project into JavaScript: sbt example/fastLinkJS 2. Open your client configuration file, such as the Cline settings file. 3. Add the server entry to the mcpServers object with stdio transport: json { "mcpServers": { "mcpscala": { "disabled": false, "timeout": 30, "command": "sh", "args": ["/path/to/run.sh"], "transportType": "stdio" } } } 4. Update the path to point to your actual runner script and restart your client.

What you can do with MCP-scala

• Exposing Scala 3 backend functions and calculations directly as tools for MCP-compatible AI agents. - Automatically deriving JSON schemas from typed Scala case classes using Tapir and Circe integrations. - Generating random numbers within bounded intervals for automated workflows directly through the randomNumber tool. - Producing sequences of integers and calculating sums using the built-in iota and sum demo tools. - Building custom stdio-based MCP servers using Cats Effect IO and the dev.capslock::mcpscala dependency.

Key facts

• Open Source
• https://github.com/windymelt/mcp-scala
• AI & LLM Tooling
• framework, jvm, scala

Part of MCP Servers

What is MCP-scala?

MCP-scala is an open source Model Context Protocol server implementation written in Scala 3. It allows Scala and JVM developers to build MCP servers, define tools, and connect them to AI clients using the stdio transport protocol.

What can MCP-scala do?

MCP-scala enables automatic derivation of JSON Schema from Scala case classes, custom tool definitions using Cats Effect IO, and text content messaging. It also includes demo tools such as randomNumber, iota, and sum for sequence operations.

Which MCP clients work with MCP-scala?

MCP-scala works with clients that support standard stdio MCP servers, including Cline. The project documentation specifically provides an example configuration for integrating the compiled server directly with the Cline interface.

Is MCP-scala open source?

Yes, MCP-scala is open source. You can access the source code, inspect example implementations, open issues, and contribute pull requests directly on its GitHub repository.

What state of development is MCP-scala in?

The project is currently in an alpha stage. Core features such as JSON Schema derivation, tool definitions, and stdio transport are functional, while HTTP transport, capability handling, and notification handling are planned for future updates.