TokenMetre

A VS Code extension that turns raw Claude Code session logs into a rich, interactive token usage dashboard — per question, per model, per hour, per day, and across your entire history.

---

Why This Exists

Claude Code stores every session as a JSONL file in ~/.claude/projects/. Those files contain precise token counts, model names, and timestamps for every API call — but there's no built-in way to actually read them.

This extension parses those files and gives you:

• Exactly how many tokens each question cost you
• Which model answered it and how many times it was called
• An estimated dollar cost using live Anthropic pricing
• Aggregates rolled up by hour, day, month, and year across all your projects

---

Features

Sessions Sidebar

A tree view of every Claude Code session, organised by project.

SESSIONS
├── my-api-service
│   ├── 2026-05-29  e3f1a2b4
│   └── 2026-05-28  9c82d051
└── frontend-app
    └── 2026-05-27  4b71e390

Click any session to open its full token report in a webview panel.

---

Per-Turn Session Report

Every question you asked in a session gets its own row, showing exactly what it cost.

#	Question	Calls	Input	Cache+	Cache~	Output	% Eff	Cost
1	refactor the auth middleware	8	42.1K	18.3K	120.4K	6.2K	14.2%	$0.052
2	add unit tests for the parser	5	38.7K	0	98.1K	4.9K	11.8%	$0.038

Each row includes:

• A progress bar showing that turn's share of total effective input
• Model chips showing which models answered and how many times (sonnet-4-6 ×8)
• Timestamp of when you asked the question

Token columns explained:

• Input — fresh, uncached tokens sent to the model
• Cache+ — tokens written to the prompt cache (billed at 1.25× input)
• Cache~ — tokens read from the prompt cache (billed at 0.1× input)
• Output — tokens generated by the model
• % Eff — this turn's share of total effective (billable) input

---

Hourly Breakdown

Within each session report, turns are grouped by the hour they occurred — useful for understanding how your work was spread across a long session.

Hour	Turns	API Calls	Eff. Input	Output	Cost
2026-05-29 09:00	4	18	280K	32K	$0.18
2026-05-29 11:00	7	31	510K	58K	$0.34
2026-05-29 14:00	3	12	190K	21K	$0.12

---

Model Breakdown (Per Session)

See which models were used in a session and exactly what each cost you.

Model	Calls	Input	Cache+	Cache~	Output	Cost
claude-sonnet-4-6	42	580K	210K	1.4M	98K	$3.18
claude-haiku-4-5	8	42K	0	88K	12K	$0.09

---

Token Usage Sidebar — Daily / Monthly / Yearly Aggregates

A second panel aggregates token usage across all sessions, all projects and groups it by time period.

TOKEN USAGE
├── Daily  (31)
│   ├── 2026-05-29   4 sess · 6.2M tok · $18.40   ▶
│   │   ├── sonnet-4-6   ×180 · 5.8M tok · $17.20
│   │   └── haiku-4-5    ×22  · 400K tok  · $1.20
│   ├── 2026-05-28   2 sess · 3.1M tok · $9.20    ▶
│   └── ...
├── Monthly  (3)
│   ├── 2026-05   18 sess · 42M tok · $124.80     ▶
│   └── ...
└── Yearly  (1)
    └── 2026   18 sess · 42M tok · $124.80        ▶

Each period row shows: session count · total tokens · estimated cost.
Expand any period to see the per-model breakdown directly in the tree.

---

Period Report Webview

Click any date (daily, monthly, or yearly row) to open a dedicated report for that period — a full table of every session that falls within it.

#	Date	Project / Session	Turns	Calls	Tokens	Output	Cost
1	2026-05-29	my-api-service · e3f1a2b4	12	48	6.2M	520K	$18.40
2	2026-05-29	frontend-app · 4b71e390	5	19	980K	88K	$2.90

Includes a model breakdown aggregated across all sessions in the period, and an Export CSV button.

---

Cost Estimation

All cost figures use Anthropic's published pricing ($/million tokens):

Model family	Input	Output	Cache write	Cache read
Opus	$15	$75	$18.75	$1.50
Sonnet	$3	$15	$3.75	$0.30
Haiku	$0.80	$4	$1.00	$0.08

Unrecognised model IDs fall back to family matching (opus / haiku / sonnet default).

---

CSV Export

Every report — session, period — has an Export CSV button. The exported file includes:

• Per-turn rows with timestamp, question text, all token columns, and cost
• An hourly breakdown section (session reports)
• A model breakdown section with per-model costs

---

Installation

This extension is not yet published to the VS Code Marketplace. To run it locally:

cd claude-tokens-extension
npm install

Then open the folder in VS Code and press F5 (or run Run Extension from the Run panel). A new Extension Development Host window will open with the extension active.

To compile without launching:

npm run compile

---

Requirements

• VS Code 1.85+
• Claude Code installed and used at least once (so ~/.claude/projects/ exists)
• Node.js 18+ (for development)

---

How It Works

Claude Code writes every session event — user messages, assistant responses, tool calls, permission prompts — as a newline-delimited JSON record to ~/.claude/projects/<encoded-path>/<session-id>.jsonl.

Each assistant response record carries a usage object with input_tokens, output_tokens, cache_creation_input_tokens, and cache_read_input_tokens. The extension:

1. Reads all JSONL files in ~/.claude/projects/
2. Identifies real human turns (filters out tool-result messages and meta expansions)
3. Walks each assistant entry's parentUuid chain to find its originating human turn
4. Deduplicates by message.id so multi-entry streaming responses are counted once
5. Aggregates token counts and model usage into per-turn buckets
6. Rolls those buckets up into hourly, daily, monthly, and yearly maps

---

License

MIT — see LICENSE.