We’re publishing the /telemetry page now, before any event is actually collected. The point: lock the commitment in writing first, then ship the implementation against it. Nothing on that page is being recorded today — telemetry is still opt-in across every FerrLabs product. The page is marked draft in a banner so nobody mistakes it for the live state.
Why this order. The honest version of “we collect telemetry” is to publish the full event list, the hashing function, the retention window, and the opt-out paths before flipping the default. We’re using the page as a contract: when on-by-default ships, it will match this document line for line, or the document changes first. The work is tracked in FerrLabs-Cloud#166.
What it covers. Per-product event tables (FerrFlow CLI, the FerrLabs API, app.ferrlabs.com, FerrVault / FerrTrack / FerrGrowth / FerrFleet), the never-collected list, BLAKE3-keyed hashing with 90-day salt rotation, TimescaleDB retention windows, and three opt-out surfaces (CLI env var, app preferences, org-wide setting). Available in EN and FR, linked from the footer of every FerrLabs site.
Until step 7 lands the principle №04 wording on ferrlabs.com still reads “No telemetry by default. Off until you opt in.” That sentence flips to “Telemetry, but transparent” only when the page is wired to read events from Kit source at build time and the opt-out works end to end.