A paid remote MCP that connects Claude to a hosted eval dashboard for AI SDK workflows. It exposes four tools: run_eval_gate for gating decisions, summarize_trace_failures for debugging, compare_eval_versions for A/B analysis, and issue_eval_receipt for audit trails. Requires a bearer token from the product site. Useful when you need structured eval verdicts and usage logs that persist outside your local environment, or when compliance teams want receipts and audit evidence for AI agent decisions. Built on streamable HTTP transport, so it works with MCP clients that support remote servers. Pricing and tokens live at futureagievals.clauxel.com.
claude mcp add --transport http com.clauxel.futureagievals-futureagievals-mcp https://futureagievals.clauxel.com/mcp