mcp · scorecard-ai — Claude Code MCP server: Scorecard API in your editor
Connects an AI assistant to the Scorecard REST API so it can browse endpoints, fire test requests and help wire the SDK into your app.
claude mcp add --transport http mcp https://mcp.scorecard.io/mcpWhat mcp does
The server sits between your assistant and Scorecard's API. The agent pulls the documentation from there, tries endpoints with real calls, and suggests how to use scorecard-ai in your code. Handy when you evaluate and tune LLM systems and don't want to copy schemas by hand.
You can reach it over HTTP at https://mcp.scorecard.io/mcp or run it locally through npx -y scorecard-ai-mcp. The local path needs SCORECARD_API_KEY; without it the server can't authenticate, so requests fail. The library itself is a regular npm package.
It works with the same objects the SDK exposes: testsets, testcases, runs. If you already have a key and a Scorecard project, adding the server to your client takes a couple of commands.
What it helps with
- browse Scorecard API endpoints from a chat window
- send a test request without hand-writing curl
- add
scorecard-aito an app with the assistant's guidance - inspect a testset or run structure using the docs
- check an API response before writing integration code
Example prompts for the agent
- List the Scorecard endpoints for working with testsets
- Send a test request to the Scorecard API and show me the response
- Help me call
client.runs.createfromscorecard-aiin my project
How to install mcp
1Platform CLI: remote server
claude mcp add --transport http mcp https://mcp.scorecard.io/mcp2Platform CLI
claude mcp add --env SCORECARD_API_KEY=<SCORECARD_API_KEY> --transport stdio scorecard-ai-mcp -- npx -y scorecard-ai-mcpReplace <…> with your own values
1Platform CLI: remote server
codex mcp add scorecard-ai-mcp --url https://mcp.scorecard.io/mcp2Platform CLI
codex mcp add scorecard-ai-mcp --env SCORECARD_API_KEY=<SCORECARD_API_KEY> -- npx -y scorecard-ai-mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"scorecard-ai-mcp": {
"url": "https://mcp.scorecard.io/mcp"
}
}
}{
"mcpServers": {
"scorecard-ai-mcp": {
"command": "npx",
"args": [
"-y",
"scorecard-ai-mcp"
],
"env": {
"SCORECARD_API_KEY": "<SCORECARD_API_KEY>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"scorecard-ai-mcp": {
"type": "http",
"url": "https://mcp.scorecard.io/mcp"
}
}
}{
"servers": {
"scorecard-ai-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"scorecard-ai-mcp"
],
"env": {
"SCORECARD_API_KEY": "<SCORECARD_API_KEY>"
}
}
}
}Replace <…> with your own values
1Platform CLI: remote server
gemini mcp add -t http scorecard-ai-mcp https://mcp.scorecard.io/mcp2Platform CLI
gemini mcp add -e SCORECARD_API_KEY=<SCORECARD_API_KEY> scorecard-ai-mcp npx -y scorecard-ai-mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about mcp
How do I install mcp in Claude Code?
For HTTP transport run claude mcp add --transport http scorecard-ai-mcp https://mcp.scorecard.io/mcp. For a local process run claude mcp add --env SCORECARD_API_KEY=<SCORECARD_API_KEY> --transport stdio scorecard-ai-mcp -- npx -y scorecard-ai-mcp.
What does mcp do?
It gives your assistant access to the Scorecard REST API: endpoint discovery, test calls, and the documentation it needs to help integrate the scorecard-ai SDK into your application.
What does the server need to work?
A SCORECARD_API_KEY. With the stdio setup you pass it via --env or your MCP client's environment variables, since API calls won't authenticate without it.
Similar MCP servers
All MCP servers for AI and ML →mcp by other authors comes first.
