Dot Token Dashboard: Claude Code Hub Usage Rankings
Summarize Token usage and model rankings from a self-hosted Claude Code Hub, then display them on Quote/0 through the Canvas API.
Dot Token Dashboard reads Token statistics from the request logs of a self-hosted Claude Code Hub and builds a 296 × 152 Canvas dashboard for Quote/0. It can show total usage for a selected period, today's usage, and a per-model ranking, then push the result to the device on a schedule.
Features
- Two display modes:
usageshows today's and period totals;completedshows the period total and full model ranking - Per-model statistics: Retrieves available models, calculates Token usage for each, and supports excluding unwanted models
- Native Canvas output: Builds a Dot Canvas node tree without rendering an intermediate PNG
- Flexible configuration: Controls the time range, timezone, labels, row count, and layout dimensions
- Local preview: Uses
--dry-runto generatepayload.jsonfor preview throughdemo.html - Scheduled delivery: Can refresh Quote/0 at a fixed interval through cron
Requirements
- Python 3 and the project dependencies
- An accessible self-hosted Claude Code Hub instance and API Key
- A Dot API Key, Quote/0 device ID, and Task Key for the target Canvas content
- Local configuration created from
config.yaml.exampleand.env.example
Provide secrets, the Claude Code Hub address, device ID, and Task Key through .env or environment variables. Use config.yaml for the reporting period, API paths, labels, and layout.
APIs and credentials
This project depends on the Claude Code Hub statistics, model-list, and
leaderboard endpoints. Update the configuration or client if their paths or
response formats change. Keep the Claude Code Hub and Dot API Keys private,
and do not commit .env, config.yaml, or a generated payload.json
containing real links and Task Keys.
Did this solve your problem?
Join our community