Cost-xray: Seeing What My AI Coding Agent Actually Sends to the API

Описание и применение

cost-xray позволяет разработчикам видеть, какие части запросов AI‑кодировщиков расходуют токены и деньги, предоставляя детальный анализ и возможности оптимизации.

Задача

Developers lack visibility into the detailed cost drivers of AI coding agents, seeing only total session cost without knowing which parts of the request consume tokens and money.

Результат

cost-xray provides fine-grained visibility into AI coding agent request traffic, enabling developers to pinpoint cost drivers, optimize prompts and tool usage, and reduce unnecessary token consumption.

Рабочий процесс

• Install cost-xray via one-line installer
• Select target agent (Claude Code, Codex, or both) during setup
• Run your coding agent normally (e.g., codex or claude)
• After coding, launch the dashboard with cx
• Navigate the dashboard: agent → project → session → category → MCP server → tool → per-turn call
• Review detailed views to see context-window occupancy and per-source cost
• Use insights to optimize agent configuration (remove unused schemas, adjust caching, limit tool outputs)

Доказательства

cost-xray is a local tool for inspecting AI coding-agent requests.

Источник

The capture layer uses mitmproxy as a local hop. It records redacted request and response data under ~/.cost-xray/.

Источник

After that, open a new terminal and use your agent normally: codex or claude, then open the dashboard: cx

Источник