Rundida MCP

What XWeaponX7/rundida-mcp MCP server does

XWeaponX7/rundida-mcp MCP server connects MCP-compatible assistants to RunDida’s running resources. Its ten tools cover two main areas: retrieving reference data from RunDida and performing running calculations locally.

The data tools expose the available running calculators, guides, and marathon events. An agent can list resources, then request detailed information about a selected calculator, guide, or event. Tool and guide details may include FAQs and related resources. Marathon details include event information, dates, locations, countdown data, and the event data described by RunDida.

The calculation tools handle common training and race questions:

• calculate_pace derives pace, finish time, or distance when any two of those values are supplied.
• predict_race estimates a target-distance result from a known race result using the Riegel formula, with VO2max estimation also described in the project documentation.
• heart_rate_zones calculates five training zones with the Karvonen method.
• marathon_countdown returns the remaining time for a selected marathon event.

How it works

XWeaponX7/rundida-mcp MCP server separates remote data retrieval from local computation. Data tools fetch content through the RunDida API and cache responses for 30 minutes. These operations therefore require an internet connection, but the project states that the API does not require authentication.

Compute tools run locally and do not make API calls. The documented formulas are the Riegel formula for race prediction, the Jack Daniels method for VO2max estimation in race prediction, and the Karvonen method for heart-rate zones. Results depend on the inputs supplied by the calling assistant, such as distance, time, pace, age, resting heart rate, or a previous race result.

The server is intended for MCP clients that can launch a local Node.js process over standard input and output. The README identifies Claude, Cursor, Windsurf, and other MCP clients as supported usage contexts.

Setup and configuration

XWeaponX7/rundida-mcp MCP server requires Node.js 18 or newer. Install and launch it through the published npm package with:

npx -y @rundida/mcp-server

For Claude Desktop, add a server entry whose command is npx and whose arguments are -y and @rundida/mcp-server. Cursor and Windsurf use the same command and argument pattern in their MCP configuration. Claude Code can register the same process with its MCP add command.

No environment variables, API keys, account, or additional service credentials are documented as required. Data features need network access to reach RunDida, while local calculation features can operate without API requests.

Tools and capabilities

The server exposes these tools:

• list_tools: browse the available running calculators.
• get_tool: retrieve details for a specific calculator, including FAQs and related tools.
• list_guides: browse RunDida’s running guides.
• get_guide: retrieve guide details, FAQs, and related tools.
• list_marathons: list tracked marathon events with dates and locations.
• get_marathon: retrieve details for a selected marathon.
• calculate_pace: calculate pace, time, or distance from two supplied values.
• predict_race: estimate a race result from a known performance.
• heart_rate_zones: calculate five Karvonen training zones.
• marathon_countdown: calculate a countdown to a selected event.

RunDida’s content is documented as available in English, Chinese, and Japanese in the package introduction. The broader RunDida platform also describes free calculators, guides, marathon information, and a JSON API.

Limitations and notes

The server does not provide general-purpose training-plan generation beyond the tools and RunDida content exposed through MCP. Remote content can be unavailable without an internet connection, and cached data may remain unchanged for up to 30 minutes. Formula-based outputs are estimates based on the supplied inputs; race predictions should not be treated as guarantees of performance. The project is distributed under the MIT license.