bench

Overview ¶

Command bench runs the mcpproxy benchmark.

Default (offline) mode scores the committed Spec 065 frozen corpus for token reduction and writes a JSON report plus a static HTML dashboard:

go run ./bench/cmd/bench [-corpus PATH] [-out DIR] [-encoding NAME]

Live mode boots against a running proxy (see bench/docker-compose.yml) to add the exact-token comparison (full schemas), retrieval accuracy (Recall@k / MRR / nDCG over the golden set), and search latency:

go run ./bench/cmd/bench -live [-proxy URL] [-api-key KEY] [-golden PATH]

Reports land in bench/results/ (gitignored — reports are never committed, per the Spec 065 CN-003 repo rule).