Your facts stay yours.
How the public scoring beta handles the information you choose to share.
Updated November 4, 2026.
What goes to the scoring service
When you use the optional connection, Grok sends the supplied resume text, extracted facts, job description and requirement rubric to Stitch Your Pitch. Contact details are unnecessary. Redact them before sharing when they are not relevant.
The backend validates exact source references and calculates a provisional heuristic score. It does not call an external model provider or retrieve job URLs during scoring.
What the backend stores
The reviewed backend processes resume and job inputs in request memory. It has no profile or result database, shared resume cache or resume hashes in quota records.
Quota storage contains only HMAC-keyed counters, window timestamps and expiry. It contains no raw account identifiers or career text. Quota records are eligible for asynchronous cleanup 48 hours after their window; expiration is also enforced by the application.
Grok, Auth0 and infrastructure
Grok handles conversations, attachments and tool results under your platform settings. Auth0 handles login and identity records under its policies. These are separate from the scoring backend.
The deployed service disables application logging and excludes Cloud Run request logs from its default logging sink. The default application-log bucket has seven-day retention; required cloud audit logs remain. This is not a promise of zero provider telemetry or immediate deletion everywhere.
Your choices
You can disconnect the scoring connection in Grok. Manage account and conversation deletion through the relevant provider controls. Do not send passwords, tokens, government identifiers, banking details or unnecessary personal data.
The website has no resume upload, contact form, analytics script or browser storage. Independent fresh-user and automatic renewal verification are still in progress during this beta.