alexei-led/analyzing-usage

<!-- Pi platform guidance --> <!-- Use Pi tool names exactly: read, bash, edit, write, ask_user_question, structured_output, todo, Agent, get_subagent_result, steer_subagent, web_search, web_answer, web_research. --> <!-- Use Agent, get_subagent_result, and steer_subagent for delegated work. --> <!-- Use ctx7 or npx ctx7@latest through bash when Context7 documentation lookup is required. -->

Analyze Pi-Agent Usage

Analyze pi-agent session usage via ccusage-pi.

• Reference: ccusage-pi flags and JSON shapes

Tool

Binary: ccusage-pi — reads ~/.pi/agent/sessions/ (override: PI_AGENT_DIR env var)

npx @ccusage/pi@latest <subcommand> [flags]

Subcommands

• daily — token/cost by date
• monthly — token/cost by month
• session — token/cost per conversation

No blocks, weekly, or statusline subcommands.

Flags

All subcommands share:

• --since YYYYMMDD — start date (also accepts YYYY-MM-DD)
• --until YYYYMMDD — end date
• --json — JSON output (required for jq piping)
• --breakdown / -b — per-model breakdown rows
• --order asc|desc — sort order (default: desc)
• --timezone / -z — IANA timezone
• --pi-path <path> — override sessions directory

No --jq, --offline, --mode, --instances, or --project flags.

Computing Dates

# macOS
SINCE=$(date -v-14d +%Y%m%d)
UNTIL=$(date +%Y%m%d)
TODAY=$(date +%Y%m%d)

# Linux
SINCE=$(date -d '14 days ago' +%Y%m%d)
UNTIL=$(date +%Y%m%d)

Detect OS before computing: uname -s returns Darwin (macOS) or Linux.

Output Rules

• Only report numbers from actual ccusage-pi output. Never invent totals.
• When result is empty: "No pi-agent usage data for this period."
• Every report must include: date range queried, total cost, top model used.
• Compute totalTokens as inputTokens + outputTokens + cacheCreationTokens + cacheReadTokens — no totalTokens field in JSON.
• Empty session list returns bare [], not { "sessions": [] } — guard with if type == "array".
• Give concrete optimization suggestions: reduce prompt size, increase cache reuse, check session reuse patterns.
• Always append 2>/dev/null to suppress npm resolver stderr.
• Prefer uvx termgraph for visualization when available; fall back to plain jq table otherwise.

Key Queries

Daily Cost (14d)

npx @ccusage/pi@latest daily --since $SINCE --until $UNTIL --json 2>/dev/null \
  | jq -r '.daily[] | "\(.date[5:]),\(.totalCost | . * 100 | round / 100)"' \
  | uvx termgraph --title "Daily Cost (USD)" --color cyan --width 60 --suffix " $"

Monthly Totals

npx @ccusage/pi@latest monthly --json 2>/dev/null \
  | jq -r '.monthly[] | "\(.month),\(.totalCost | . * 100 | round / 100)"' \
  | uvx termgraph --title "Monthly Cost (USD)" --color green --width 60 --suffix " $"

Session Breakdown (today)

npx @ccusage/pi@latest session --since $TODAY --until $TODAY --json 2>/dev/null \
  | jq -r 'if type == "array" then .[] else .sessions[] end
    | "\(.sessionId | split("-") | last),\(.totalCost | . * 100 | round / 100)"'

Cache Efficiency

npx @ccusage/pi@latest daily --since $SINCE --until $UNTIL --json 2>/dev/null \
  | jq -r '.daily[] |
    (.cacheReadTokens + .cacheCreationTokens) as $cache |
    (.inputTokens + .outputTokens + $cache) as $total |
    select($total > 0) |
    "\(.date[5:]),\($cache / $total * 100 | . * 10 | round / 10)"' \
  | uvx termgraph --title "Cache Rate %" --color blue --width 60 --suffix "%"

Token Volume

npx @ccusage/pi@latest daily --since $SINCE --until $UNTIL --json 2>/dev/null \
  | jq -r '.daily[] |
    (.inputTokens + .outputTokens + .cacheCreationTokens + .cacheReadTokens) as $total |
    "\(.date[5:]),\($total / 1000000 | . * 100 | round / 100)"' \
  | uvx termgraph --title "Daily Tokens (M)" --color yellow --width 60 --suffix " M"

What to Show

• Default: daily cost + cache efficiency (14d window)
• --today: session breakdown + cache efficiency (7d context)
• monthly: monthly cost chart
• Full analysis: all five queries above