Real-Time Dashboard & Cost Analytics

My Claude gives you a live dashboard that shows exactly where your Claude Code usage stands - not after the fact, but right now. Utilization gauges, burn rate, cost breakdowns, and pace predictions all in one view.

Dashboard Overview

My Claude dashboard showing usage metrics, charts, and cost analytics

The main dashboard displays:

5-Hour Window Gauge - Your current utilization in the rolling 5-hour window, color-coded green/yellow/red
7-Day Window Gauge - Your utilization in the rolling 7-day allocation
7-Day Sonnet Window - Separate gauge for Sonnet-specific rate limits
Total Cost - Estimated or API-synced costs for the selected period
Total Sessions - Number of Claude Code sessions
Total Messages - Message count across all sessions

Burn Rate

My Claude calculates your current burn rate in real-time:

Tokens per hour - How fast you're consuming tokens
Cost per hour - What your current pace is costing you
Messages per hour - Your interaction velocity

This tells you at a glance whether your current work pace is sustainable or if you need to slow down.

Pace Predictions

The pace system projects your usage forward and tells you exactly when you'll hit your limit:

On Track (green) - Current usage rate is sustainable, you'll reset before exhaustion
Warning (yellow) - You may hit the limit if you continue at this rate
Critical (red) - You will hit the limit before the window resets, with the exact predicted time

Instead of a vague "you're at 73%", you see "You'll hit your cap at 3:15pm." That's the difference between a surprise lockout and a deliberate decision to pace yourself.

Charts and Analytics

Activity Heatmap

A hour-by-hour heatmap showing your usage intensity across the day. Immediately see your peak coding hours and when you tend to burn through tokens the fastest.

Daily Activity Chart

Area chart showing messages and tokens over time with dual-axis visualization. Track trends across days and weeks.

Model Cost Breakdown

Stacked bar chart breaking down costs by model (Opus, Sonnet, Haiku) and token type:

Input tokens (blue) - What you sent to Claude
Output tokens (purple) - What Claude generated
Cache read tokens (green) - Free tokens from prompt caching
Cache write tokens (amber) - Initial cache population cost

Cache Savings Visualization

Cost analysis with cache savings comparison

See exactly how much prompt caching is saving you:

Actual cost vs what you would have paid without caching
Total saved with percentage (e.g., "Saved $47.23 (42.3%)")

This alone can justify your plan choice and show the ROI of structuring your prompts for cache hits.

Context Window Breakdown

See exactly what's consuming your context window in the current session:

System prompt allocation
Memory/CLAUDE.md content
Conversation history
Free space remaining
Autocompact buffer size

This is unique to My Claude - most tools don't show you which part of your prompt is eating tokens.

Per-Project Tracking

My Claude discovers your projects from session data and provides per-project breakdowns. Compare KPIs side-by-side across projects to see which codebases drive the most usage and cost.

Data Sources

Usage data comes from two sources:

Local JSONL files - Session transcripts from Claude Code (immediate, real-time)
Anthropic API - Billing data and rate limit status (authoritative for costs)

When API data is available, it takes precedence over local calculations. The data source is shown for each metric so you always know what you're looking at.