Vibe Coding Tracker — AI Coding Assistant Usage Tracker

Track your AI coding costs in real-time. Vibe Coding Tracker is a powerful CLI tool that helps you monitor and analyze your Claude Code, Codex, and Gemini usage, providing detailed cost breakdowns, token statistics, and code operation insights.

English | 繁體中文 | 简体中文

Note: CLI examples use the short alias vct. If you built from source, the compiled binary is named vibe_coding_tracker; create an alias or replace vct with the full name when running commands.

🎯 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.

📊 Beautiful Visualizations

Choose your preferred view:

Interactive Dashboard: Auto-refreshing terminal UI with live updates
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, and Gemini. No setup required—just run and analyze.

🎨 Rich Insights

Token usage by model and date
Cost breakdown by cache types
File operations tracking
Command execution history
Git repository information

✨ Key Features

Feature	Description
🤖 Auto-Detection	Intelligently identifies Claude Code, Codex, or Gemini logs
💵 Smart Pricing	Fuzzy model matching + daily cache for speed
🎨 4 Display Modes	Interactive, Table, Text, and JSON outputs
📈 Comprehensive Stats	Tokens, costs, file ops, and tool calls
⚡ High Performance	Built with Rust for speed and reliability
🔄 Live Updates	Real-time dashboard refreshes every second
💾 Efficient Caching	Smart daily cache reduces API calls

🚀 Quick Start

Installation

Prerequisites:

Rust toolchain (1.70+) for building from source
Node.js (v18+) for npm installation

From `npm` (Recommended)

npm install -g vibe-coding-tracker

From Source

# Clone and build
git clone https://github.com/Mai0313/VibeCodingTracker.git
cd VibeCodingTracker
cargo build --release

# Binary location
./target/release/vibe_coding_tracker

# Optional: create a short alias (adjust the paths as needed)
ln -sf "$(pwd)/target/release/vibe_coding_tracker" ~/.local/bin/vct

First Run

# View your usage with the short alias (if available)
vct usage

# Or run the binary built by Cargo
./target/release/vibe_coding_tracker usage

# Analyze a specific conversation
./target/release/vibe_coding_tracker analysis --path ~/.claude/projects/session.jsonl

💡 Tip: Use vct as a short alias for vibe_coding_tracker to save typing—create it manually with ln -sf ./target/release/vibe_coding_tracker ~/.local/bin/vct (or any path you prefer).

📖 Command Guide

🔍 Quick Reference

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

Commands:
usage       Show token usage and costs (default: interactive)
analysis    Analyze conversation files and export data
version     Display version information
update      Update to the latest version from GitHub releases
help        Show help information

💰 Usage Command

Track your spending across all AI coding sessions.

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
vct usage --json

What You Get

The tool scans these directories automatically:

~/.claude/projects/*.jsonl (Claude Code)
~/.codex/sessions/*.jsonl (Codex)
~/.gemini/tmp/<project_hash>/chats/*.json (Gemini)

🎨 Interactive Mode (Default)

Live dashboard that updates every second

┌──────────────────────────────────────────────────────────────────┐
│                  📊 Token Usage Statistics                       │
└──────────────────────────────────────────────────────────────────┘
┌────────────┬──────────────────────┬────────────┬────────────┬────────────┬──────────────┬────────────┬────────────┐
│ Date       │ Model                │ Input      │ Output     │ Cache Read │ Cache Create │ Total      │ Cost (USD) │
├────────────┼──────────────────────┼────────────┼────────────┼────────────┼──────────────┼────────────┼────────────┤
│ 2025-10-01 │ claude-sonnet-4-20…  │ 45,230     │ 12,450     │ 230,500    │ 50,000       │ 338,180    │ $2.15      │
│ 2025-10-02 │ claude-sonnet-4-20…  │ 32,100     │ 8,920      │ 180,000    │ 30,000       │ 251,020    │ $1.58      │
│ 2025-10-03 │ claude-sonnet-4-20…  │ 28,500     │ 7,200      │ 150,000    │ 25,000       │ 210,700    │ $1.32      │
│ 2025-10-03 │ gpt-4-turbo          │ 15,000     │ 5,000      │ 0          │ 0            │ 20,000     │ $0.25      │
│            │ TOTAL                │ 120,830    │ 33,570     │ 560,500    │ 105,000      │ 819,900    │ $5.30      │
└────────────┴──────────────────────┴────────────┴────────────┴────────────┴──────────────┴────────────┴────────────┘
┌──────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│ 💰 Total Cost: $5.30  |  🔢 Total Tokens: 819,900  |  📅 Entries: 4  |  🧠 Memory: 12.5 MB                       │
└──────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘

Press 'q', 'Esc', or 'Ctrl+C' to quit

Features:

✨ Auto-refreshes every second
🎯 Highlights today's entries
🔄 Shows recently updated rows
💾 Displays memory usage
📊 Summary statistics

Controls: Press q, Esc, or Ctrl+C to exit

📋 Static Table Mode

Perfect for documentation and reports

vct usage --table

📊 Token Usage Statistics

╔════════════╦══════════════════════╦════════════╦════════════╦════════════╦══════════════╦══════════════╦════════════╗
║ Date       ║ Model                ║ Input      ║ Output     ║ Cache Read ║ Cache Create ║ Total Tokens ║ Cost (USD) ║
╠════════════╬══════════════════════╬════════════╬════════════╬════════════╬══════════════╬══════════════╬════════════╣
║ 2025-10-01 ║ claude-sonnet-4-20…  ║ 45,230     ║ 12,450     ║ 230,500    ║ 50,000       ║ 338,180      ║ $2.15      ║
║ 2025-10-02 ║ claude-sonnet-4-20…  ║ 32,100     ║ 8,920      ║ 180,000    ║ 30,000       ║ 251,020      ║ $1.58      ║
║ 2025-10-03 ║ claude-sonnet-4-20…  ║ 28,500     ║ 7,200      ║ 150,000    ║ 25,000       ║ 210,700      ║ $1.32      ║
║            ║ TOTAL                ║ 105,830    ║ 28,570     ║ 560,500    ║ 105,000      ║ 799,900      ║ $5.05      ║
╚════════════╩══════════════════════╩════════════╩════════════╩════════════╩══════════════╩══════════════╩════════════╝

📝 Text Mode

Ideal for scripting and parsing

vct usage --text

2025-10-01 > claude-sonnet-4-20250514: $2.154230
2025-10-02 > claude-sonnet-4-20250514: $1.583450
2025-10-03 > claude-sonnet-4-20250514: $1.321200
2025-10-03 > gpt-4-turbo: $0.250000

🗂️ JSON Mode

Full precision for accounting and integration

vct usage --json

{
  "2025-10-01": [
    {
      "model": "claude-sonnet-4-20250514",
      "usage": {
        "input_tokens": 45230,
        "output_tokens": 12450,
        "cache_read_input_tokens": 230500,
        "cache_creation_input_tokens": 50000,
        "cache_creation": {
          "ephemeral_5m_input_tokens": 50000
        },
        "service_tier": "standard"
      },
      "cost_usd": 2.1542304567890125
    }
  ]
}

🔍 Output Comparison

Feature	Interactive	Table	Text	JSON
Best For	Monitoring	Reports	Scripts	Integration
Cost Format	$2.15	$2.15	$2.154230	2.1542304567890123
Updates	Real-time	Static	Static	Static
Colors	✅	✅	❌	❌
Parseable	❌	❌	✅	✅

💡 Use Cases

Budget Tracking: Monitor your daily AI spending
Cost Optimization: Identify expensive sessions
Team Reporting: Generate usage reports for management
Billing: Export precise costs for invoicing
Monitoring: Real-time dashboard for active development

📊 Analysis Command

Deep dive into conversation files - single file or batch analysis.

Basic Usage

# Single file: Analyze and display
vct analysis --path ~/.claude/projects/session.jsonl

# Single file: Save to file
vct analysis --path ~/.claude/projects/session.jsonl --output report.json

# Batch: Analyze all sessions with interactive table (default)
vct analysis

# Batch: Save aggregated results to JSON
vct analysis --output batch_report.json

# Batch with provider grouping: Output complete records grouped by provider (JSON format)
vct analysis --all

# Save the grouped results to a file
vct analysis --all --output grouped_report.json

What You Get

Single File Analysis:

Token Usage: Input, output, and cache statistics by model
File Operations: Every read, write, and edit with full details
Command History: All shell commands executed
Tool Usage: Counts of each tool type used
Metadata: User, machine ID, Git repo, timestamps

Batch Analysis:

Aggregated Metrics: Grouped by date and model
Line Counts: Edit, read, and write operations
Tool Statistics: Bash, Edit, Read, TodoWrite, Write counts
Interactive Display: Real-time TUI table (default)
JSON Export: Structured data for further processing

Sample Output - Single File

{
  "extensionName": "Claude-Code",
  "insightsVersion": "0.1.0",
  "user": "wei",
  "machineId": "5b0dfa41ada84d5180a514698f67bd80",
  "records": [
    {
      "conversationUsage": {
        "claude-sonnet-4-20250514": {
          "input_tokens": 252,
          "output_tokens": 3921,
          "cache_read_input_tokens": 1298818,
          "cache_creation_input_tokens": 124169
        }
      },
      "toolCallCounts": {
        "Read": 15,
        "Write": 4,
        "Edit": 2,
        "Bash": 5,
        "TodoWrite": 3
      },
      "totalUniqueFiles": 8,
      "totalWriteLines": 80,
      "totalReadLines": 120,
      "folderPath": "/home/wei/repo/project",
      "gitRemoteUrl": "https://github.com/user/project.git"
    }
  ]
}

Sample Output - Batch Analysis

Interactive Table (default when running vct analysis):

┌──────────────────────────────────────────────────────────────────┐
│                  🔍 Analysis Statistics                           │
└──────────────────────────────────────────────────────────────────┘
┌────────────┬────────────────────┬────────────┬────────────┬────────────┬──────┬──────┬──────┬───────────┬───────┐
│ Date       │ Model              │ Edit Lines │ Read Lines │ Write Lines│ Bash │ Edit │ Read │ TodoWrite │ Write │
├────────────┼────────────────────┼────────────┼────────────┼────────────┼──────┼──────┼──────┼───────────┼───────┤
│ 2025-10-02 │ claude-sonnet-4-5…│ 901        │ 11,525     │ 53         │ 13   │ 26   │ 27   │ 10        │ 1     │
│ 2025-10-03 │ claude-sonnet-4-5…│ 574        │ 10,057     │ 1,415      │ 53   │ 87   │ 78   │ 30        │ 8     │
│ 2025-10-03 │ gpt-5-codex        │ 0          │ 1,323      │ 0          │ 75   │ 0    │ 20   │ 0         │ 0     │
│            │ TOTAL              │ 1,475      │ 22,905     │ 1,468      │ 141  │ 113  │ 125  │ 40        │ 9     │
└────────────┴────────────────────┴────────────┴────────────┴────────────┴──────┴──────┴──────┴───────────┴───────┘
┌────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│ 📝 Total Lines: 25,848  |  🔧 Total Tools: 428  |  📅 Entries: 3  |  🧠 Memory: 8.2 MB                         │
└────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘

Press 'q', 'Esc', or 'Ctrl+C' to quit

JSON Export (with --output):

[
  {
    "date": "2025-10-02",
    "model": "claude-sonnet-4-5-20250929",
    "editLines": 901,
    "readLines": 11525,
    "writeLines": 53,
    "bashCount": 13,
    "editCount": 26,
    "readCount": 27,
    "todoWriteCount": 10,
    "writeCount": 1
  },
  {
    "date": "2025-10-03",
    "model": "claude-sonnet-4-5-20250929",
    "editLines": 574,
    "readLines": 10057,
    "writeLines": 1415,
    "bashCount": 53,
    "editCount": 87,
    "readCount": 78,
    "todoWriteCount": 30,
    "writeCount": 8
  }
]

💡 Use Cases

Single File Analysis:

Usage Auditing: Track what the AI did in each session
Cost Attribution: Calculate costs per project or feature
Compliance: Export detailed activity logs
Analysis: Understand coding patterns and tool usage

Batch Analysis:

Productivity Tracking: Monitor coding activity over time
Tool Usage Patterns: Identify most-used tools across sessions
Model Comparison: Compare efficiency between different AI models
Historical Analysis: Track trends in code operations by date

🔧 Version Command

Check your installation.

# Formatted output
vct version

# JSON format
vct version --json

# Plain text
vct version --text

Output

🚀 Vibe Coding Tracker

╔════════════════╦═════════╗
║ Version        ║ 0.1.0   ║
╠════════════════╬═════════╣
║ Rust Version   ║ 1.89.0  ║
╠════════════════╬═════════╣
║ Cargo Version  ║ 1.89.0  ║
╚════════════════╩═════════╝

🔄 Update Command

Keep your installation up-to-date automatically.

The update command checks GitHub releases and downloads the latest version for your platform.

Basic Usage

# Interactive update with confirmation
vct update

# Check for updates without installing
vct update --check

# Force update without confirmation prompt
vct update --force

How It Works

Check Latest Version: Fetches the latest release from GitHub API
Compare Versions: Compares current version with the latest available
Download Binary: Downloads the appropriate binary for your platform (Linux/macOS/Windows)
Smart Replacement:
- Linux/macOS: Automatically replaces the binary (backs up old version to .old)
- Windows: Downloads as .new and creates a batch script for safe replacement

Platform Support

The update command automatically detects your platform and downloads the correct archive:

Linux: vibe_coding_tracker-v{version}-linux-x64-gnu.tar.gz, vibe_coding_tracker-v{version}-linux-arm64-gnu.tar.gz
macOS: vibe_coding_tracker-v{version}-macos-x64.tar.gz, vibe_coding_tracker-v{version}-macos-arm64.tar.gz
Windows: vibe_coding_tracker-v{version}-windows-x64.zip, vibe_coding_tracker-v{version}-windows-arm64.zip

Windows Update Process

On Windows, the binary cannot be replaced while running. The update command:

Downloads the new version as vct.new
Creates an update script (update_vct.bat)
Displays instructions to complete the update

Run the batch script after closing the application to finish the update.

💡 Smart Pricing System

How It Works

Automatic Updates: Fetches pricing from LiteLLM daily
Smart Caching: Stores pricing in ~/.vibe-coding-tracker/ for 24 hours
Fuzzy Matching: Finds best match even for custom model names
Always Accurate: Ensures you get the latest pricing

Model Matching

Priority Order:

✅ Exact Match: claude-sonnet-4 → claude-sonnet-4
🔄 Normalized: claude-sonnet-4-20250514 → claude-sonnet-4
🔍 Substring: custom-gpt-4 → gpt-4
🎯 Fuzzy (AI-powered): Uses Jaro-Winkler similarity (70% threshold)
💵 Fallback: Shows $0.00 if no match found

Cost Calculation

Total Cost = (Input Tokens × Input Cost) +
             (Output Tokens × Output Cost) +
             (Cache Read × Cache Read Cost) +
             (Cache Creation × Cache Creation Cost)

🐳 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 ~/.gemini:/root/.gemini \
    vibe_coding_tracker:latest usage

🔍 Troubleshooting

Pricing Data Not Loading

# Check cache
ls -la ~/.vibe-coding-tracker/

# Force refresh
rm -rf ~/.vibe-coding-tracker/
vct usage

# Debug mode
RUST_LOG=debug vct usage

No Usage Data Shown

# Verify session directories
ls -la ~/.claude/projects/
ls -la ~/.codex/sessions/
ls -la ~/.gemini/tmp/

# Count session files
find ~/.claude/projects -name "*.jsonl" | wc -l
find ~/.codex/sessions -name "*.jsonl" | wc -l
find ~/.gemini/tmp -name "*.json" | wc -l

Analysis Command Fails

# Validate JSONL format
jq empty < your-file.jsonl

# Check file permissions
ls -la your-file.jsonl

# Run with debug output
RUST_LOG=debug vct analysis --path your-file.jsonl

Interactive Mode Issues

# Reset terminal if broken
reset

# Check terminal type
echo $TERM  # Should be xterm-256color or compatible

# Use static table as fallback
vct usage --table

⚡ Performance

Built with Rust for speed and reliability:

Operation	Time
Parse 10MB JSONL	~320ms
Analyze 1000 events	~45ms
Load cached pricing	~2ms
Interactive refresh	~30ms

Binary Size: ~3-5 MB (stripped)

📚 Learn More

Developer Docs: See .github/copilot-instructions.md
Report Issues: GitHub Issues
Source Code: GitHub Repository

🤝 Contributing

Contributions welcome! Here's how:

Fork the repository
Create your feature branch
Make your changes
Submit a pull request

For development setup and guidelines, see .github/copilot-instructions.md.

📄 License

MIT License - see LICENSE for details.

🙏 Credits

LiteLLM for model pricing data
Claude Code, Codex, and Gemini teams for creating amazing AI coding assistants
The Rust community for excellent tooling

Save money. Track usage. Code smarter.

⭐ Star this project if you find it useful!

Made with 🦀 Rust

vibe_coding_tracker 0.2.0

Vibe Coding Tracker — AI Coding Assistant Usage Tracker

🎯 Why Vibe Coding Tracker?

💰 Know Your Costs

📊 Beautiful Visualizations

🚀 Zero Configuration

🎨 Rich Insights

✨ Key Features

🚀 Quick Start

Installation

From npm (Recommended)

From Source

First Run

📖 Command Guide

🔍 Quick Reference

💰 Usage Command

Basic Usage

What You Get

🎨 Interactive Mode (Default)

📋 Static Table Mode

📝 Text Mode

🗂️ JSON Mode

🔍 Output Comparison

💡 Use Cases

📊 Analysis Command

Basic Usage

What You Get

Sample Output - Single File

Sample Output - Batch Analysis

💡 Use Cases

🔧 Version Command

Output

🔄 Update Command

Basic Usage

How It Works

Platform Support

Windows Update Process

💡 Smart Pricing System

How It Works

Model Matching

Cost Calculation

🐳 Docker Support

🔍 Troubleshooting

Pricing Data Not Loading

No Usage Data Shown

Analysis Command Fails

Interactive Mode Issues

⚡ Performance

📚 Learn More

🤝 Contributing

📄 License

🙏 Credits

From `npm` (Recommended)