Speedrun logo

Speedrun

Searches Speedrun.com for games, categories, leaderboards, and users through MCP.

Install Speedrun

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "speedrun": {
      "command": "npx",
      "args": [
        "-y",
        "@pipeworx/mcp-speedrun"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

Speedrun MCP server connects AI agents to the Speedrun.com API v1 for searching games, inspecting categories, retrieving ranked runs, and finding users. Agents can use the hosted Pipeworx gateway over streamable HTTP or run the package locally as a stdio MCP server. The gateway endpoint also exposes shared Pipeworx tools, while the standalone package provides only the Speedrun tools. Reach for it when an agent needs structured speedrunning records, category IDs, player profiles, or links to Speedrun.com pages.

Use cases

  • Search for a game on Speedrun.com
  • List categories for a selected game
  • Compare top runs in a leaderboard
  • Find a speedrunner profile by username

Key features

  • Searches games by name
  • Lists game categories and IDs
  • Retrieves top-N ranked runs
  • Finds users by username
  • Supports hosted HTTP and local stdio connections

Tools4

What Speedrun lets your agent do. Expand a tool for its parameters.

Self-reported~241 tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 1d ago

Recent health check succeeded.

7/7 checks healthy over the last 39d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 13d ago via OSV.dev · @pipeworx/mcp-speedrun (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

npm downloads398Package downloads in the last 30 days.
Last commit26d agoMost recent push to the default branch.
Tools exposed4Callable tools this server registers over MCP.
Directory activity1 viewsConfig copies, upvotes, and views on AllMCPs.

How Speedrun works

What the Speedrun MCP server does

The Speedrun MCP server exposes selected Speedrun.com API v1 data to MCP-compatible clients. It supports a lookup workflow that starts with a game search, continues by retrieving that game's categories, and then uses the resulting identifiers to fetch a leaderboard. A separate user search supports finding speedrunners by username.

Returned game records include the game ID, abbreviation, release year, and Speedrun.com link. Category results include category IDs and names such as Any%, 100%, or Glitchless. Leaderboard results contain ranked runs, durations in ISO-8601 format and seconds, run dates, player information, and run links. User results include matching IDs, display names, and profile links.

How it works

The tools use identifiers from earlier calls where the API workflow requires them:

  • search_games accepts a game-name query and an optional maximum result count.
  • get_categories accepts the game ID returned by search_games.
  • get_leaderboard accepts a game ID, category ID, and optional top-result count.
  • find_user searches for a speedrunner by username.

For example, an agent can search for “Super Mario 64,” select a matching game, list its categories, and request the top 10 runs for a selected category. The results provide links for checking the corresponding records on Speedrun.com.

Setup and configuration

The Speedrun MCP server can be connected through the hosted Pipeworx endpoint:

config.json
{
  "mcpServers": {
    "speedrun": {
      "url": "https://gateway.pipeworx.io/speedrun/mcp"
    }
  }
}

Claude Desktop, Cursor, and Windsurf are named as supported client examples. The hosted connection does not require an account for its first calls. Its tool list includes the four Speedrun tools plus shared Pipeworx meta-tools, including routing and discovery functions.

For a local connection, run the package with npx:

config.json
{
  "mcpServers": {
    "speedrun": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-speedrun"]
    }
  }
}

The local process communicates over stdin and stdout and does not require a Pipeworx account or a gateway round trip. It lists only the Speedrun package tools.

Limitations and notes

The Speedrun MCP server is focused on read operations described by the available tools: game search, category lookup, leaderboard retrieval, and user lookup.

The hosted pack endpoint and the full Pipeworx gateway use the same gateway infrastructure and data-source catalog, but they differ in which pack tools are listed directly. The pack endpoint includes shared gateway meta-tools by design. For a connection containing only the Speedrun tools, use the standalone stdio package instead.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Speedrun

Speedrun is an MCP server that connects AI agents to the Speedrun.com API v1. Its main tools search games, list game categories, retrieve top-ranked leaderboards, and find users by username. Results include identifiers, links, run times, dates, and player details needed for researching speedrunning data.

Run it locally with `npx -y @pipeworx/mcp-speedrun` and configure the command and argument in your MCP client. The package communicates over stdin and stdout, requires no Pipeworx account, and exposes only the Speedrun tools.

Yes, Cursor is listed as a supported client example. Add the hosted `https://gateway.pipeworx.io/speedrun/mcp` URL to Cursor's MCP configuration, or configure the local `npx` stdio command.

Yes, use the standalone npm package locally without a Pipeworx account. The hosted gateway also requires no account for its first calls, but its connection includes shared Pipeworx meta-tools in addition to the Speedrun tools.

First call `search_games` to obtain a game ID, then call `get_categories` with that ID to obtain a category ID. Pass both IDs to `get_leaderboard`, optionally providing `top` to request a chosen number of ranked runs.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewSpeedrun AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/speedrun?style=directory)](https://allmcps.com/mcp/speedrun)
HTML Embed
<a href="https://allmcps.com/mcp/speedrun"><img src="https://allmcps.com/api/badge/speedrun?style=directory" alt="Speedrun on AllMCPs" /></a>