telemetry.danimoya.com

Anonymous install-count receiver for Claude-Dashboard, Claude-B, and related projects. Opt-in, off by default.

What this is

Each dashboard install can voluntarily POST four fields to /v1/ping once a week:

{ "installation_id":  "",
  "dashboard_version": "1.4.2",
  "heliosdb_version":  "3.19.1",
  "timestamp":         "2026-05-03T12:34:56Z" }

The receiver hashes (weekly_salt, client_ip, installation_id) and stores only the hash + version columns. Raw IP and the raw (ip, id) tuple are never persisted. Salt rotates weekly so hashes from week N can't be cross-correlated with hashes from week N+1.

Public stats

Open /stats for a per-week chart of reported installs and cumulative ping-events. JSON for the same numbers is available at /v1/stats/history and /v1/stats/cumulative.

Endpoints

POST /v1/ping — submit a ping (called by the dashboards)
GET /v1/stats — current week aggregate, public
GET /v1/stats/history?weeks=N — last N weeks of installs (N ≤ 104), public
GET /v1/stats/cumulative?weeks=N — same series with a running sum, public
GET /stats — SVG-rendered chart of the above

Source

Receiver source, schema, salt-rotation cron, and append-only audit changelog: github.com/danimoya/telemetry.

Per-install transparency

Each running dashboard exposes its own /telemetry page showing exactly what that install would send. Read that page on your own deployment before deciding whether to opt in.