llm-cost-tracker

Track and report LLM token usage and cost for OpenClaw via OpenRouter.

Quick Start

python3 scripts/collect_usage.py --init      # first-time: create DB + backfill
python3 scripts/run_tracker.py               # telegram report (default)
python3 scripts/run_tracker.py --output terminal  # full terminal report
python3 scripts/run_tracker.py --output debug     # per-request debug

Core Design

• Source of truth: usage.cost.total from OpenRouter — never recomputed from token counts
• Data source: OpenClaw session JSONL files (including .reset files from context compaction)
• Deduplication: by openrouter_request_id — safe to re-run backfill anytime
• Time windows: 24h rolling (UTC); 7d/30d/90d/365d calendar days (configurable timezone)

Schema: request_facts

| Column | Source | |--------|--------| | openrouter_request_id | responseId (unique key) | | created_at_utc | entry timestamp | | model | model ID | | prompt_tokens | usage.input | | completion_tokens | usage.output | | cached_tokens | usage.cacheRead | | cache_write_tokens | usage.cacheWrite | | reasoning_tokens | usage.reasoning | | total_tokens | usage.totalTokens | | billed_cost | usage.cost.total (canonical) |

Configuration

config/env.json (all fields optional):

{
  "SESSIONS_DIR": "",
  "TIMEZONE": "Asia/Hong_Kong",
  "UTC_OFFSET_HOURS": 8
}

• SESSIONS_DIR: override session file location (empty = auto-detect)
• TIMEZONE / UTC_OFFSET_HOURS: for calendar-day window calculations
• API key: auto-detected from ~/.openclaw/agents/main/agent/auth-profiles.json or OPENROUTER_API_KEY env var

Scheduled Reports

openclaw cron add --name "llm-cost:collect" --message "collect usage data" \
  --cron "5 0 * * *" --tz "Asia/Hong_Kong" --session isolated --no-deliver

openclaw cron add --name "llm-cost:daily" --message "llm cost" \
  --cron "0 9 * * *" --tz "Asia/Hong_Kong" --session isolated