parsimony-cli 0.2.2

Token usage and cost observability for Claude Code sessions

Token usage and cost observability for Claude Code

---

You use Claude Code daily but have no idea which sessions cost the most, which tools burn tokens, or whether caching is helping. Parsimony answers all of these. No API keys needed. It reads the JSONL files Claude Code already saves on your machine.

v0.2.0 adds filtering, budget alerts, session diffs, cost trends, and a live terminal dashboard.

---

Example Output

---

Install

pip install parsimony-cli

For the live dashboard:

pip install parsimony-cli[dashboard]

Other methods

pipx install parsimony-cli              # isolated install
pipx install parsimony-cli[dashboard]   # with live dashboard
uv tool install parsimony-cli           # with uv
python -m parsimony                     # if not in PATH

---

Usage

parsimony                # today's summary (default)
parsimony yesterday      # yesterday's report
parsimony week           # this week
parsimony week --last    # last week
parsimony month          # this month
parsimony month 2026-03  # specific month

Filtering

Narrow any report to specific models, tools, or cost ranges:

parsimony week --model sonnet                    # only Sonnet sessions
parsimony today --model opus --model haiku        # Opus or Haiku
parsimony month --tool Read --tool Write          # sessions using Read or Write
parsimony week --min-cost 0.50                    # sessions costing $0.50+
parsimony top sessions --max-cost 1.00            # cheap sessions only
parsimony week --model sonnet --min-cost 0.10     # combine filters

Live Dashboard

Real-time terminal dashboard that auto-refreshes as Claude Code sessions generate new data. Requires the dashboard extras.

parsimony live                    # launch dashboard
parsimony live --project myapp    # filter by project

Key	Action
q	Quit
r	Force refresh
t	Toggle today/week/month

The dashboard shows: cost summary with trend arrow, per-model cost bars, top tools by call count, cache hit rate gauge, and a scrollable session log.

Budget Alerts

Set daily, weekly, or monthly cost budgets. Warnings appear automatically in matching reports when you approach or exceed limits.

parsimony budget    # view current budget status

Configure in ~/.parsimony/config.yaml:

budget:
  daily: 5.00
  weekly: 25.00
  monthly: 80.00

Budget status shows as OK / NOTE (70%+) / WARN (90%+) / OVER for each configured period.

Cost Trends

Visualize cost over time with daily bars, 7-day moving averages, and automatic trend direction detection:

parsimony trend              # 30-day cost trend (default)
parsimony trend --days 7     # last 7 days
parsimony trend --days 90    # last 90 days

Session Diff

Compare two sessions side-by-side to see how workflow changes affect cost:

parsimony diff a1b2c3d4 e5f6a7b8    # compare by prefix or full UUID

Shows deltas for total cost, tokens, cache efficiency, per-model breakdown, and per-tool usage with color-coded arrows.

Session Drill-Down

parsimony session a1b2c3d4   # prefix match or full UUID

Rankings

parsimony top sessions --period week    # most expensive sessions
parsimony top models   --period month   # cost by model
parsimony top tools    --period all     # most used tools
parsimony top projects --period week    # cost by project

Compare Periods

parsimony compare --period week  --last 4   # last 4 weeks side-by-side
parsimony compare --period month --last 3   # last 3 months

Export

parsimony --export json month > report.json
parsimony --export csv week > models.csv
parsimony trend --days 7 --export json > trend.json
parsimony diff a1b2 c3d4 --export json > diff.json

---

How It Works

graph LR
    A["~/.claude/projects/*.jsonl"] --> B["Scanner"]
    B --> C["Reader"]
    C --> D["Session Builder"]
    D --> E["Cost Engine"]
    D --> F["Grouper"]
    D --> L["Filters"]
    E --> G["Rollup"]
    F --> G
    L --> G
    G --> H["Rich Tables & Charts"]
    G --> I["JSON / CSV Export"]
    G --> M["Trends & Budget"]
    H --> J["Terminal"]
    I --> K["File"]
    M --> H
    A --> N["Watcher"]
    N --> O["Live Dashboard"]
    G --> O

    style A fill:#1e3a5f,stroke:#22d3ee,color:#22d3ee
    style B fill:#1a1a2e,stroke:#8b949e,color:#e6edf3
    style C fill:#1a1a2e,stroke:#8b949e,color:#e6edf3
    style D fill:#1a1a2e,stroke:#8b949e,color:#e6edf3
    style E fill:#1a1a2e,stroke:#e879f9,color:#e879f9
    style F fill:#1a1a2e,stroke:#e879f9,color:#e879f9
    style G fill:#1a1a2e,stroke:#4ade80,color:#4ade80
    style H fill:#1a1a2e,stroke:#22d3ee,color:#22d3ee
    style I fill:#1a1a2e,stroke:#22d3ee,color:#22d3ee
    style J fill:#1e3a5f,stroke:#22d3ee,color:#22d3ee
    style K fill:#1e3a5f,stroke:#22d3ee,color:#22d3ee
    style L fill:#1a1a2e,stroke:#e879f9,color:#e879f9
    style M fill:#1a1a2e,stroke:#4ade80,color:#4ade80
    style N fill:#1a1a2e,stroke:#facc15,color:#facc15
    style O fill:#1e3a5f,stroke:#facc15,color:#facc15

Data Pipeline

graph TD
    subgraph Parse
        A["JSONL line-by-line"] --> B["Dedup by requestId"]
        B --> C["Detect model switches"]
        C --> D["Extract tool usage"]
    end

    subgraph Calculate
        D --> E["Decimal-precision costs"]
        E --> F["Group by model / tool / project / day"]
        F --> G["Cache efficiency metrics"]
        G --> G2["Filters / Trends / Budgets"]
    end

    subgraph Output
        G2 --> H["Rich tables"]
        G2 --> I["Bar charts & gauges"]
        G2 --> J["JSON / CSV"]
        G2 --> K["Live Dashboard (Textual)"]
    end

    style A fill:#0d1117,stroke:#22d3ee,color:#22d3ee
    style B fill:#0d1117,stroke:#22d3ee,color:#22d3ee
    style C fill:#0d1117,stroke:#22d3ee,color:#22d3ee
    style D fill:#0d1117,stroke:#22d3ee,color:#22d3ee
    style E fill:#0d1117,stroke:#e879f9,color:#e879f9
    style F fill:#0d1117,stroke:#e879f9,color:#e879f9
    style G fill:#0d1117,stroke:#e879f9,color:#e879f9
    style G2 fill:#0d1117,stroke:#e879f9,color:#e879f9
    style H fill:#0d1117,stroke:#4ade80,color:#4ade80
    style I fill:#0d1117,stroke:#4ade80,color:#4ade80
    style J fill:#0d1117,stroke:#4ade80,color:#4ade80
    style K fill:#0d1117,stroke:#facc15,color:#facc15

What Each Report Shows

Section	Details
Summary	Total cost, session count, API calls
By Model	Per-model tokens, cost, share %
By Tool	Tool call counts, MCP vs built-in
Cache	Hit rate gauge, read/write breakdown
Sessions	Time, duration, project, model, cost
Budget	Spend vs limit with OK/WARN/OVER status
Trends	Daily cost bars, 7-day moving average, direction
Diff	Side-by-side session comparison with deltas
Dashboard	All of the above, live-updating in real time

---

Pricing

Built-in pricing for all Claude models. Override at ~/.parsimony/pricing.yaml:

Default pricing table

Model	Input	Output	Cache Write	Cache Read
Opus 4.6	$5.00/M	$25.00/M	$6.25/M	$0.50/M
Sonnet 4.6	$3.00/M	$15.00/M	$3.75/M	$0.30/M
Haiku 4.5	$1.00/M	$5.00/M	$1.25/M	$0.10/M

Unknown models fall back to Sonnet pricing.

---

Project Structure

graph TD
    CLI["cli.py - Click entry point"] --> Parser
    CLI --> Models
    CLI --> Aggregator
    CLI --> Output
    CLI --> Cache
    CLI --> Dashboard
    CLI --> Budget

    subgraph Parser
        P1["scanner.py - discover files"]
        P2["reader.py - stream JSONL"]
        P3["session_builder.py - dedup & segments"]
        P4["events.py - frozen dataclasses"]
    end

    subgraph Models
        M1["session.py - domain model"]
        M2["cost.py - Decimal engine"]
        M3["tool_usage.py - MCP parsing"]
    end

    subgraph Aggregator
        A1["time_range.py - time windows"]
        A2["grouper.py - group by dimension"]
        A3["rollup.py - full aggregation"]
        A4["filters.py - model/tool/cost filters"]
        A5["trends.py - moving averages & direction"]
        A6["diff.py - session comparison"]
    end

    subgraph Output
        O1["tables.py - Rich tables"]
        O2["charts.py - bar charts & gauge"]
        O3["formatters.py - $1.23, 1.2M"]
        O4["export.py - JSON & CSV"]
        O5["diff_table.py - diff renderer"]
    end

    subgraph Dashboard
        D1["app.py - Textual TUI"]
        D2["widgets.py - live-updating panels"]
        D3["watcher.py - filesystem monitor"]
    end

    subgraph Budget
        B1["budget.py - thresholds & alerts"]
    end

    subgraph Cache
        C1["store.py - SQLite cache"]
    end

    style CLI fill:#1e3a5f,stroke:#22d3ee,color:#22d3ee
    style Parser fill:#0d1117,stroke:#8b949e,color:#e6edf3
    style Models fill:#0d1117,stroke:#e879f9,color:#e6edf3
    style Aggregator fill:#0d1117,stroke:#4ade80,color:#e6edf3
    style Output fill:#0d1117,stroke:#facc15,color:#e6edf3
    style Dashboard fill:#0d1117,stroke:#f97316,color:#e6edf3
    style Budget fill:#0d1117,stroke:#f43f5e,color:#e6edf3
    style Cache fill:#0d1117,stroke:#60a5fa,color:#e6edf3

---

Contributing

git clone https://github.com/MinaSaad1/parsimony.git
cd parsimony
pip install -e ".[dev,dashboard]"
pytest                # 260 tests, 80%+ coverage
ruff check src/       # lint
mypy src/             # type check

---

License

MIT License. See LICENSE for details.