Author verifiable eval records through a draft → review → revise → submit loop with server-enforced graders; compile to JSONL/CSV/Inspect/lm-eval via MCP. STDIO or Streamable HTTP.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add evals -- npx -y @cyanheads/evals-mcp-server~/.codex/config.toml
[mcp_servers.evals]
command = "npx"
args = ["-y", "@cyanheads/evals-mcp-server"]
# codex mcp add evals -- npx -y @cyanheads/evals-mcp-server~/.cursor/mcp.json
{
"mcpServers": {
"evals": {
"command": "npx",
"args": [
"-y",
"@cyanheads/evals-mcp-server"
]
}
}
}.vscode/mcp.json
{
"servers": {
"evals": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@cyanheads/evals-mcp-server"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"evals": {
"command": "npx",
"args": [
"-y",
"@cyanheads/evals-mcp-server"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"evals": {
"command": "npx",
"args": [
"-y",
"@cyanheads/evals-mcp-server"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"evals": {
"command": "npx",
"args": [
"-y",
"@cyanheads/evals-mcp-server"
]
}
}
}