Claude Code Statusline with 1M Context Support

Dynamic context window detection for Claude Code statusline that automatically adjusts to your model's context window size.

Features

• 🎯 Dynamic Context Detection: Automatically detects [1m], [200k], etc. suffixes
• 📊 Accurate Percentages: Shows correct context usage for 1M and 200K models
• 🔢 Token Count: Shows absolute usage next to the percentage (e.g. 25% 250K/1M)
• 💰 Session Metrics: Cost tracking and duration monitoring
• 💸 Premium Pricing Alert: Shows "💸2x" only for legacy Sonnet 4/4.5 above 200K tokens (current models have flat 1M pricing)
• 🎨 Color-coded Alerts: Visual warnings at 50%, 75%, 90%, 95% usage
• 😱 Keyboard Layout Indicator (macOS): screaming-face emoji when a non-English layout is active — catches "ааа, не та раскладка!" before you type

Installation

🚀 Automatic Installation (Recommended)

Run the installer - it will set everything up for you:

npx claude-statusline-1m --install

The installer will:

1. Ask where you want to install (global, project, or local)
2. Copy the Python script to the right location
3. Update your settings.json automatically
4. Show you what was installed

Quick install with defaults (installs locally without prompts):

npx claude-statusline-1m --install --yes

Then restart Claude Code and you're done! 🎉

🛠️ Manual Installation (Advanced)

If you prefer to configure manually, add this to your .claude/settings.json or .claude/settings.local.json:

For npm version:

{
  "statusLine": {
    "type": "command",
    "command": "npx -y claude-statusline-1m"
  }
}

For GitHub version:

{
  "statusLine": {
    "type": "command",
    "command": "npx -y github:vedmichv/claude-statusline-1m"
  }
}

Then restart Claude Code.

How It Works

• 🚀 NPX downloads and caches the package automatically
• ⚡ Claude Code runs it on every session
• 📦 No installation script needed - runs directly from cache
• 🌍 Works at project OR global level
• 🔄 Always uses the cached version for fast startup

Quick Test

Test the statusline before configuring:

# Test with npm package
echo '{"model":{"id":"test[1m]","display_name":"Claude"},"workspace":{"current_dir":"/tmp"},"transcript_path":""}' | npx -y claude-statusline-1m

# Test with GitHub version
echo '{"model":{"id":"test[1m]","display_name":"Claude"},"workspace":{"current_dir":"/tmp"},"transcript_path":""}' | npx -y github:vedmichv/claude-statusline-1m

Expected output: [Claude] 📁 tmp 🧠 🔵 ???

Context Window Detection

The script automatically detects your model's context window from the model ID:

• model[1m] → 1,000,000 tokens
• model[200k] → 200,000 tokens
• No suffix → 200,000 tokens (default)

Supported Models

• ✅ Claude Fable 5 / Mythos 5 ([1m])
• ✅ Claude Sonnet 5 ([1m])
• ✅ Claude Opus 4.6 / 4.7 / 4.8 ([1m])
• ✅ Claude Sonnet 4.6 ([1m])
• ✅ Legacy Claude Sonnet 4 / 4.5 with 1M context ([1m]) — with 💸2x premium alert
• ✅ Claude Haiku 4.5 (200K)
• ✅ Any future models with [XM] or [Xk] suffixes

Pricing Information (Amazon Bedrock, July 2026)

| Model | Input | Output | >200K surcharge | |-------|-------|--------|-----------------| | Fable 5 / Mythos 5 | $10/M | $50/M | None (flat 1M) | | Sonnet 5 | $2/M intro → $3/M | $10/M intro → $15/M | None (flat 1M) | | Opus 4.8 / 4.7 / 4.6 | $5/M | $25/M | None (flat 1M) | | Sonnet 4.6 | $3/M | $15/M | None (flat 1M) | | Haiku 4.5 | $1/M | $5/M | — (200K window) | | Legacy Sonnet 4 / 4.5 on the old 1M beta | $3/M → $6/M | $15/M → $22.50/M | 2x input / 1.5x output above 200K (grandfathered only) |

Sonnet 5 introductory pricing ($2/$10) runs through August 31, 2026. Regional (non-global) Bedrock endpoints cost +10% over the prices above.

⚠️ Legacy note: the Sonnet 4/4.5 1M-context beta appears retired — the current Bedrock pricing page no longer lists the >200K tier, and Anthropic docs list Sonnet 4.5 as a 200K model again. The surcharge can only still apply to grandfathered sessions that run Sonnet < 4.6 with a [1m] model ID. For those, when you exceed 200K tokens, ALL tokens in that request are charged at the premium rate, not just the excess.

The statusline displays a 💸2x indicator only in exactly that case (Sonnet < 4.6 on a 1M window above 200K tokens) — a conservative alert for grandfathered sessions. Current models (Fable 5, Sonnet 5, Opus 4.x, Sonnet 4.6) have flat pricing across the full 1M window, so no indicator is shown.

Requirements

• Python 3.x
• Claude Code CLI

License

MIT