Multi-PC Aggregation

Multi-PC aggregation is claude-telemetry's core differentiator. This page explains how it works and how to get the most out of it.

Why Multi-PC Matters

Anthropic's Claude Code rate limits are per account, not per machine. If you use Claude Code on your work laptop and home desktop, your combined usage counts against the same limits.

Without aggregation:

• You can't see total spending across machines
• You might hit rate limits unexpectedly because you didn't account for usage on another PC
• There's no way to know which machine is consuming the most tokens

How It Works

1. Each PC runs its own agent — the agent calls ccusage to extract usage data from local logs
2. Agents push to Supabase — each record is tagged with a machine_id
3. Dashboard aggregates — queries combine data from all machines

The agent does no custom JSONL parsing — it delegates that to ccusage and focuses only on multi-PC aggregation and centralized sync.

Machine IDs

Each agent generates a unique UUID on first run, stored in ~/.claude-tracker/config.json. This ID is immutable — it stays the same even if you rename the machine.

The machine_name field is a human-friendly label (e.g., "work-laptop", "home-desktop") that you set during setup. You can change it at any time.

Dashboard Filtering

Every dashboard page supports machine filtering:

• All Machines — the default view, shows aggregated data
• Single Machine — filter by a specific machine to see its isolated usage
• Compare — side-by-side comparison between two machines

The machine selector appears in the top navigation bar on every page.

Adding a New Machine

1. Clone the repo and set up the agent (see Agent Setup)
2. Run claude-tracker setup with the same Supabase credentials
3. Give it a unique machine_name
4. Run claude-tracker install-service

The new machine will appear in the dashboard within 15 minutes (the default sync interval).

Removing a Machine

To stop syncing from a machine:

claude-tracker uninstall-service
claude-tracker uninstall

Historical data from that machine remains in Supabase. To remove it:

DELETE FROM usage_data WHERE machine_id = 'your-machine-uuid';

Sync Behavior

• Interval: Every 15 minutes by default (configurable via claude-tracker daemon --interval N)
• Idempotent: The agent tracks the last synced timestamp — re-syncing doesn't create duplicates
• Force re-sync: Use claude-tracker sync --force to re-sync all data from scratch
• Offline resilient: If the network is down, the agent retries on the next interval
• Lightweight: Only new sessions since the last sync are transmitted