cost-track
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseCost Track
成本追踪
Reads the active Claude Code session jsonl (), tallies assistant-message per model, computes USD cost using REFERENCE.md pricing, and writes a structured record to the AgentDB namespace. This is the producer that gives and real data to consume.
~/.claude/projects/<encoded-cwd>/<session>.jsonlusagecost-trackingcost-reportcost-optimize读取活跃的Claude Code会话jsonl文件(),按模型统计助手消息的数据,使用REFERENCE.md中的定价计算USD成本,并将结构化记录写入 AgentDB命名空间。这是为和提供真实数据的生产者。
~/.claude/projects/<encoded-cwd>/<session>.jsonlusagecost-trackingcost-reportcost-optimizeWhen to use
使用场景
- After a meaningful chunk of work, to capture spend for the report.
- At session-end to persist the final tally.
- Periodically during long sessions (cron-friendly — see ).
/loop 30m
- 在完成一段重要工作后,捕获支出数据用于生成报告。
- 会话结束时,持久化最终统计结果。
- 在长会话期间定期执行(支持cron定时任务——参见)。
/loop 30m
Steps
操作步骤
-
Run the tracker from the project root:bash
node plugins/ruflo-cost-tracker/scripts/track.mjsThe script auto-discovers the session jsonl from the current working directory. To pin a specific session:. To dry-run (no memory write):TRACK_SESSION=/path/to/session.jsonl.TRACK_DRY_RUN=1 -
Inspect the markdown summary — total cost, per-model and per-tier breakdowns, and the persisted memory key.
-
Verify persistence —should list the new record.
mcp__claude-flow__memory_search --query "session-" --namespace cost-trackingstep 1 reads from this namespace.cost-report
-
从项目根目录运行追踪器:bash
node plugins/ruflo-cost-tracker/scripts/track.mjs脚本会自动从当前工作目录发现会话jsonl文件。要指定特定会话:。要执行试运行(不写入内存):TRACK_SESSION=/path/to/session.jsonl。TRACK_DRY_RUN=1 -
查看markdown摘要——包含总成本、按模型和层级划分的明细,以及持久化的内存键。
-
验证持久化结果——执行应能列出新记录。
mcp__claude-flow__memory_search --query "session-" --namespace cost-tracking的第一步会从此命名空间读取数据。cost-report
Record shape (in cost-tracking
namespace)
cost-trackingcost-tracking
命名空间中的记录结构
cost-trackingKey: . Value (JSON):
session-<sessionId>json
{
"sessionId": "1dba3b8c-...",
"cwd": "/Users/cohen/Projects/ruflo",
"startedAt": "2026-05-04T...",
"endedAt": "2026-05-05T...",
"messageCount": 234,
"byModel": {
"claude-opus-4-7": {
"tier": "opus",
"input_tokens": 12345,
"output_tokens": 6789,
"cache_creation_input_tokens": 800000,
"cache_read_input_tokens": 2000000,
"messages": 50,
"cost_usd": 1.23
}
},
"byTier": { "haiku": 0.0, "sonnet": 0.45, "opus": 1.23, "unknown": 0 },
"total_cost_usd": 1.68,
"capturedAt": "2026-05-05T..."
}键:。值(JSON格式):
session-<sessionId>json
{
"sessionId": "1dba3b8c-...",
"cwd": "/Users/cohen/Projects/ruflo",
"startedAt": "2026-05-04T...",
"endedAt": "2026-05-05T...",
"messageCount": 234,
"byModel": {
"claude-opus-4-7": {
"tier": "opus",
"input_tokens": 12345,
"output_tokens": 6789,
"cache_creation_input_tokens": 800000,
"cache_read_input_tokens": 2000000,
"messages": 50,
"cost_usd": 1.23
}
},
"byTier": { "haiku": 0.0, "sonnet": 0.45, "opus": 1.23, "unknown": 0 },
"total_cost_usd": 1.68,
"capturedAt": "2026-05-05T..."
}Pricing source of truth
定价来源
The script's constant mirrors REFERENCE.md "Model pricing (USD per 1M tokens)". Update both together when prices change. Cache-write tokens are billed at rate; cache-read tokens at (per Anthropic billing docs).
PRICINGcache_writecache_read脚本中的常量与REFERENCE.md中的“Model pricing (USD per 1M tokens)”保持一致。当价格变动时,需同时更新两者。缓存写入token按费率计费;缓存读取token按费率计费(依据Anthropic计费文档)。
PRICINGcache_writecache_readEnv overrides
环境变量覆盖
| Env | Default | Purpose |
|---|---|---|
| | Override which project's sessions to scan |
| most-recent jsonl | Pin a specific session file |
| unset | Also write the JSON summary to this path |
| unset | Skip the |
| unset | Suppress markdown output |
| | Override target namespace |
| 环境变量 | 默认值 | 用途 |
|---|---|---|
| | 覆盖要扫描的项目会话目录 |
| 最新的jsonl文件 | 指定特定会话文件 |
| 未设置 | 将JSON摘要同时写入指定路径 |
| 未设置 | 跳过 |
| 未设置 | 抑制markdown输出 |
| | 覆盖目标命名空间 |
Cross-references
交叉引用
- (consumer) — reads records produced by this skill
cost-report - REFERENCE.md "Cost attribution formula" — the math the script implements
- (consumer; landing in P2) — reads totals to evaluate alerts
cost-budget-check
- (消费者)——读取本工具生成的记录
cost-report - REFERENCE.md中的“Cost attribution formula”——脚本实现的计算逻辑
- (消费者;计划在P2阶段推出)——读取总支出数据以生成告警
cost-budget-check