vibe-coding-tracker 0.15.0

Vibe Coding Tracker - AI coding assistant telemetry/usage parser, aggregate JSONL events into CodeAnalysis results

Vibe Coding Tracker — AI Coding Assistant Usage Tracker

Track your AI coding costs in real-time. Vibe Coding Tracker is a lightweight, high-performance CLI tool built in Rust that monitors and analyzes your Claude Code, Codex, Copilot, Gemini, and OpenCode usage — with detailed cost breakdowns, token statistics, and code operation insights, all while keeping the memory footprint minimal.

English | 繁體中文 | 简体中文

Note: CLI examples use the short alias vct. If you installed via npm/pip/cargo, the binary might be named vibe_coding_tracker or vct. Create an alias or replace vct with the full name when running commands if needed.

---

Why Vibe Coding Tracker?

Know Your Costs

Stop wondering how much your AI coding sessions cost. Get real-time cost tracking with automatic pricing updates from LiteLLM.

Ultra-Lightweight

Built with Rust for minimal resource footprint. The interactive TUI dashboard typically sits at under ~50 MB of resident memory once the first refresh is done, even with hundreds of long-context sessions on disk — no Electron, no bloated runtimes. The usage path parses each session file in a lean usage-only mode and bypasses the cache, and we tune glibc's arena count at startup to keep long-running RSS honest.

Beautiful Visualizations

Choose your preferred view:

• Interactive Dashboard: Auto-refreshing terminal UI with live updates, scrollable model list (arrow keys), and compact K/M/B number formatting
• Static Reports: Professional tables for documentation
• Script-Friendly: Plain text and JSON for automation
• Full Precision: Export exact costs for accounting

Zero Configuration

Automatically detects and processes logs from Claude Code, Codex, Copilot, Gemini, and OpenCode. No setup required — just run and analyze.

Rich Insights

• Token usage by model and date
• Cost breakdown by cache types (read / create)
• File operations tracking (edit, read, write lines)
• Tool call history (Bash, Edit, Read, Write, TodoWrite)
• Per-provider totals

---

Key Features

Feature	Description
Multi-Provider	Claude Code, Codex, Copilot, and Gemini — all in one place
Smart Pricing	Fuzzy model matching + daily cache from LiteLLM
4 Display Modes	Interactive TUI, static table, plain text, and JSON
Dual Analysis	Token/cost stats (usage) + code operation stats (analysis)
Ultra-Lightweight	Under ~50 MB RSS in the TUI, streaming JSONL parse — built with Rust
Live Updates	Real-time dashboard refreshes every second
Efficient Caching	Smart daily cache reduces API calls

---

Quick Start

Installation

Choose the installation method that works best for you:

Developers: If you want to build from source or contribute to development, please see CONTRIBUTING.md.

Method 1: Install from npm

Prerequisites: Node.js v22 or higher

Choose one of the following package names (they are identical):

# Main package
npm install -g vibe-coding-tracker

# Short alias with scope
npm install -g @mai0313/vct

# Full name with scope
npm install -g @mai0313/vibe-coding-tracker

Method 2: Install from PyPI

Prerequisites: Python 3.8 or higher

pip install vibe_coding_tracker
# Or with uv
uv pip install vibe_coding_tracker

Method 3: Install from crates.io

Install using Cargo from the official Rust package registry:

cargo install vibe_coding_tracker

First Run

# View your usage with the interactive dashboard
vct usage

# Or run the binary built by Cargo/pip
vibe_coding_tracker usage

# Analyze code operations across all sessions
vct analysis

---

Command Guide

Quick Reference

vct <COMMAND> [OPTIONS]
# Replace with `vibe_coding_tracker` if you are using the full binary name

Commands:
  analysis    Analyze JSONL conversation files (single file or all sessions)
  usage       Display token usage statistics
  statusline  Cache Claude Code rate limits from a statusLine hook (stdin JSON)
  version     Display version information
  update      Update to the latest version from GitHub releases
  help        Print this message or the help of the given subcommand(s)

Time range flags (shared by usage and analysis, mutually exclusive, default --all):

Flag	Window
--daily	Sessions modified today
--weekly	Current ISO week (Monday → today)
--monthly	Current calendar month
--all	Every session on disk (default)

---

Usage Command

Track your spending across all AI coding sessions.

Flags

Flag	Purpose
(none)	Interactive TUI dashboard (default)
--table	Static table, no TUI
--text	Plain text, script-friendly
--json	JSON with enriched pricing metadata
--daily / --weekly / --monthly / --all	Time range filter (see table above)

Basic Usage

# Interactive dashboard (recommended)
vct usage

# Static table for reports
vct usage --table

# Plain text for scripts
vct usage --text

# JSON for data processing (includes cost_usd and matched_model fields)
vct usage --json

# Combine time range with output format
vct usage --weekly
vct usage --table --monthly
vct usage --json --daily

[!NOTE] Model rows are sorted by cost in ascending order, so the highest-spending model is listed last (right above the TOTAL row in --table). This applies to the interactive dashboard, --table, and --text output; --json preserves the same order. The interactive dashboard also hides models with zero usage in the selected range.

Preview: Interactive Dashboard (vct usage)

┌─────────────────────────────────────────────────────────────────────────────────────────────┐
│ Model                         Input   Output  Cache Read  Cache Write    Total  Cost (USD)  │
│                                                                                             │
│ gemini-3.1-pro-preview         129K    10.3K       67.4K            0     207K       $0.40  │
│ claude-haiku-4-5-20251001     5.57K    19.8K       4.63M         620K    5.27M       $1.34  │
│ claude-opus-4-6               25.7K     179K       40.8M        2.57M    43.6M      $77.59  │
└─────────────────────────────────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────────────────────────────┐
│ Provider                      Tokens      Cost   Active Days                                │
│                                                                                             │
│ Claude Code                    48.9M    $78.93             3                                │
│ Gemini                          207K     $0.40             1                                │
└─────────────────────────────────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────────────────────────────┐
│  Total Cost: $79.33  |  Total Tokens: 49.3M  |  Models: 3  |  Memory: 42.8 MB               │
└─────────────────────────────────────────────────────────────────────────────────────────────┘
  ↑/↓ scroll  PgUp/PgDn page  g/G top/end  r refresh  q quit  |  ★ github.com/Mai0313/VibeCodingTracker

What It Scans

The tool automatically scans these directories:

• ~/.claude/projects/**/*.jsonl (Claude Code — recursive, includes subagent logs)
• ~/.codex/sessions/**/*.jsonl (Codex — recursive, includes daily subdirectories)
• ~/.copilot/session-state/<sessionId>/events.jsonl (Copilot CLI)
• ~/.gemini/tmp/<project_hash>/chats/*.jsonl (Gemini CLI)
• ~/.local/share/opencode/opencode.db (OpenCode — SQLite database; honors $XDG_DATA_HOME)

Live Quota Panels

At the bottom of the interactive dashboard, the per-provider stats sit on the left and two live quota panels sit on the right:

┌ Provider/Tokens/Cost/Days ┬ Claude ────────┬ Codex ──────────┐
│ Claude    1.2M  $3.00  4d │ 5h ▰▰▱▱▱  16%  │ Plan: plus      │
│ Codex      800K $0.00  6d │    ↻ 4h13m     │ 5h ▰▰▱▱▱  27%   │
│ ...                       │ 7d ▰▰▰▱▱  28%  │ 7d ▱▱▱▱▱   4%   │
│                           │ updated 2m ago │ Credits: 0  +2  │
└───────────────────────────┴────────────────┴─────────────────┘

• Claude — 5-hour and weekly rate-limit usage. Claude Code only exposes these limits through its statusLine hook, so wire vct statusline ingest into your statusLine (see Statusline Command) for this panel to populate.
• Codex — plan tier, 5-hour and weekly usage, and credit balance. Pulled live from the ChatGPT backend (wham/usage) using the token in ~/.codex/auth.json on a background thread; when unavailable it falls back to the newest rate_limits in your Codex session logs (the title shows (API) vs (session)).

Quota panels appear only in the interactive TUI; --table, --text, and --json are unchanged.

---

Analysis Command

Deep dive into code operations — see exactly what your AI assistant did.

Flags

Flag	Purpose
(none)	Interactive TUI dashboard over all sessions
--path <FILE>	Analyze a single JSONL/JSON conversation file (prints JSON)
--table	Static table with per-provider totals
--text	Plain text, script-friendly
--json	JSON array of aggregated rows printed to stdout
--output <FILE>	Save results as pretty-printed JSON
--daily / --weekly / --monthly / --all	Time range filter (see table above)

See examples/ for sample inputs and matching JSON outputs for all four providers.

Basic Usage

# Interactive dashboard for all sessions (default)
vct analysis

# Static table output with per-provider totals
vct analysis --table

# Plain text for scripts
vct analysis --text

# JSON of aggregated rows for data processing
vct analysis --json

# Analyze a single conversation file → stdout JSON
vct analysis --path ~/.claude/projects/session.jsonl

# Save results to JSON
vct analysis --output report.json

# Combine time range with output format
vct analysis --weekly
vct analysis --table --monthly
vct analysis --json --daily
vct analysis --output today.json --daily

Preview: Interactive Dashboard (vct analysis)

┌─────────────────────────────────────────────────────────────────────────────────────────────┐
│ Model                     Edit Lines  Read Lines Write Lines  Bash  Edit  Read  Write       │
│                                                                                             │
│ claude-haiku-4-5-20251001           0           0           0    43     0    59      0      │
│ claude-opus-4-6                1.28K       13.3K       1.58K    82   146   209     62       │
│ gemini-3.1-pro-preview             0           0           0     0     0     0      0       │
└─────────────────────────────────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────────────────────────────┐
│ Provider        Edit Lines Read Lines Write Lines Bash Edit Read TodoWrite Write Days       │
│                                                                                             │
│ Claude Code          1.28K      13.3K       1.58K  125  146  268        18    62    3       │
│ Gemini                   0          0           0    0    0    0         0     0    1       │
└─────────────────────────────────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────────────────────────────┐
│  Total Lines: 16.1K  |  Total Tools: 619  |  Models: 3  |  Memory: 41.2 MB                  │
└─────────────────────────────────────────────────────────────────────────────────────────────┘
  ↑/↓ scroll  PgUp/PgDn page  g/G top/end  r refresh  q quit  |  ★ github.com/Mai0313/VibeCodingTracker

---

Statusline Command

Feed Claude Code rate limits into the usage dashboard.

Claude Code only exposes its 5-hour / weekly rate limits through the statusLine hook's stdin JSON, so vct captures them from there and caches them to ~/.vibe_coding_tracker/claude_rate_limits.json for the Claude quota panel.

If you already have a statusLine

Add one backgrounded line to your existing statusLine script. It has zero impact on your current display and is isolated, so a vct error can never disturb the statusLine:

printf '%s' "$input" | vct statusline ingest >/dev/null 2>&1 &

If you do not have a statusLine yet

Point Claude Code's ~/.claude/settings.json straight at vct, which caches the limits and prints a compact one-line status:

{
  "statusLine": {
    "type": "command",
    "command": "vct statusline"
  }
}

---

Update Command

Keep your installation up-to-date automatically.

The update command works for all installation methods (npm/pip/cargo/manual) by directly downloading and replacing the binary from GitHub releases.

Basic Usage

# Check for updates
vct update --check

# Interactive update with confirmation
vct update

# Force update — always downloads latest version
vct update --force

Preview (vct update --check)

Current version: v0.10.3
Checking for latest release...
Latest version: v0.10.3 — you are up to date!

---

Version Command

Report the embedded build metadata (binary version, Rust toolchain, Cargo version):

vct version          # Pretty table
vct version --text   # One-field-per-line, script-friendly
vct version --json   # Machine-readable JSON

The binary version is produced at build time by build.rs from git describe, so development builds include commit count + short SHA + dirty suffix when applicable.

---

Smart Pricing System

How It Works

1. Automatic Updates: Fetches pricing from LiteLLM daily
2. Smart Caching: Stores pricing in ~/.vibe_coding_tracker/ for 24 hours
3. Fuzzy Matching: Finds best match even for custom model names
4. Always Accurate: Ensures you get the latest pricing

Model Matching

Priority Order:

1. Exact Match: claude-sonnet-4 → claude-sonnet-4
2. Normalized: claude-sonnet-4-20250514 → claude-sonnet-4
3. Substring: custom-gpt-4 → gpt-4
4. Fuzzy (AI-powered): Uses Jaro-Winkler similarity (70% threshold)
5. Fallback: Shows $0.00 if no match found

---

Docker Support

# Build image
docker build -f docker/Dockerfile --target prod -t vibe_coding_tracker:latest .

# Run with your sessions
docker run --rm \
    -v ~/.claude:/root/.claude \
    -v ~/.codex:/root/.codex \
    -v ~/.copilot:/root/.copilot \
    -v ~/.gemini:/root/.gemini \
    -v ~/.local/share/opencode:/root/.local/share/opencode \
    vibe_coding_tracker:latest usage