MCP Tool Chainer

You use the MCP Tool Chainer to build pipelines that chain multiple MCP tools, passing results from one tool to the next to reduce token usage and streamline complex data processing tasks.

How to use

You run this MCP server locally and connect to it with an MCP client. Design a chain where the output from one tool becomes the input for the next, and use explicit placeholders to pass results between steps. This lets you perform multi-step workflows without sending large intermediate payloads back to the language model.

How to install

Prerequisites you need before installing include Node.js version 16 or later and npm.

Install from npm globally or run directly with npx.

# Install
npm install @thirdstrandstudio/mcp-tool-chainer

# Or use with npx directly
npx -y @thirdstrandstudio/mcp-tool-chainer

If you prefer to install from source, clone the repository, install dependencies, and build the package.

# Clone the repository
git clone https://github.com/thirdstrandstudio/mcp-tool-chainer.git
cd mcp-tool-chainer

# Install dependencies
npm install

# Build the package
npm run build

Additional setup for integration

If you are integrating with Claude Desktop or similar clients, you configure the MCP client to load this chainer as the last step in your workflow so it can perform final tool discovery and execution.

Configuration and usage notes

The MCP Tool Chainer supports discovering tools from all configured MCP servers and can chain multiple tools in a single call. When building a pipeline, specify the tools in sequence and refer to the intermediate results with CHAIN_RESULT placeholders. You can also apply JsonPath input and output filtering to extract only the data you need between steps.

Examples of workflow patterns

• Chain a fetch tool to retrieve a page
• Use an extraction tool to pull the headline, then pass it to a summarization tool
• Store the final result or return it to your LLM for final formatting

Available tools

mcp_chain

Chains together multiple MCP servers, executing them in sequence and passing the result from one tool as input to the next using CHAIN_RESULT placeholders.

chainable_tools

Discovers tools from all MCP servers so the mcp_chain tool can be used to build pipelines.

discover_tools

Rediscovers tools from all MCP servers to refresh available options and tool endpoints.