Back to directory

inferock-bench

A local proxy that measures OpenAI, Anthropic, Gemini, and OpenRouter API traffic to produce independent, per-call receipts of token usage, failures, and billing accuracy.

inferock-bench

inferock-bench is a local diagnostic proxy for metered LLM API traffic. It routes calls through localhost using the developer's own saved provider API keys, and records usage, pricing evidence, request and response metadata, status, timing, and retry evidence for OpenAI, Anthropic, Gemini Developer API, and pinned OpenRouter endpoints.

From these records it renders independent receipts through its own grading code and a published standard, so a team can check whether a bill matches what was actually delivered, including delivery failures such as billed-empty output or refusals that can affect a bill or audit trail.

The project discloses that its maker also sells a hosted product graded by the same standard, and publishes its spec, source, and public run cards so the methodology can be checked independently.

More in Monitoring

See all 79
  • The Standup

    standup.thecompound.tech

    A daily email and dashboard tracking software releases, security advisories, outages and breaking changes across the vendors and repositories a team depends on, surfacing only what needs action rather than a full news feed.

  • An event-driven, harness-neutral API and command-line tool that watches every coding agent running on a machine and lets a developer inspect the sessions, subagents and tool calls they leave behind, without polling.

  • claudit

    github.com

    A self-hosted dashboard for visualizing Claude Code usage from session transcripts, with time-correct cost accounting, per-project attribution, and support for ingesting Codex and Kimi CLI transcripts as well.