evalpair
Level CLocalRun LLM eval suites on local or hosted models; compare runs with intervals and a noise floor.
Installation
MCP client config (paste into mcp.json)
{
"mcpServers": {
"evalpair": {
"command": "npx",
"args": [
"evalpair"
]
}
}
}Agent install prompt
Copy and paste into Cursor / Claude / other agents to install this MCP for you
Please read this MCP repository and install it into my current environment:
https://github.com/Srikarmk/evalpair
Known working client config (can write into mcp.json):
{
"mcpServers": {
"evalpair": {
"command": "npx",
"args": [
"evalpair"
]
}
}
}
Please: 1) check README for deps and env vars; 2) write config for my client; 3) if secrets are missing, tell me what to fill — do not invent values.
Related servers
datasniffWhat's in a dataset before you download it: columns, licence and size from HuggiKeystoneWhat a research paper builds on, disputes and assumes, read straight from its LaagnoThe self-driving agent platformagnoThe programming language for agentic software.12_contextThe self-improving agent platformgemini_interactionsThe self-improving agent platform