Privacy at IntroDock
IntroDock is an early service operated by Nilesh Jahagirdar. It helps visitors compose introductions for page owners, with optional personal-agent retrieval.
Visitors
Text conversation and optional voice are processed by OpenAI. IntroDock does not record raw audio or store the chat transcript. Only the name, reply email and introduction you explicitly review and share are stored in the owner’s private inbox and available to their connected agent, if any. Your email is unverified. Owners may respond through their own services, subject to those services’ policies.
Introductions are retained for 30 days from creation. While the service is running, cleanup is scheduled about every 30 minutes, including without visitor activity; it also runs at startup. Outages or cleanup failures can delay deletion. Related retrieval, inbox-view, ownership and retry records are removed with expired introductions. Audit event types are retained for 14 days. Signed visitor session cookies last one hour. Daily and short-window limits use pseudonymous identifiers. Short-lived hashed fingerprints help prevent repeated introductions. Abuse controls are necessary to run this service; there is no advertising or analytics.
Owners
Google sign-in provides a stable account identifier and verified email. IntroDock requests no Gmail or calendar permissions. We store your public profile settings, session hashes, connector credential hashes and connection/test activity, per-connection retrieval times and resumable retrieval progress, and the first time you open an introduction in your inbox. Account sessions last up to seven days. Public profiles expose only the information you publish.
You can pause or unpublish your page, revoke its connector and delete your account through the dashboard. The original pilot account requires manual deletion/migration to preserve its separate legacy connection safely. For privacy requests or pilot deletion, reach the operator through Nilesh’s IntroDock and identify the relevant receipt or account; do not include passwords or API keys.
Shared links and previews
Published pages expose the owner’s chosen name, greeting, descriptor and photo. Social sharing services can fetch a public preview without signing in and may cache it under their own policies. Changing a page address keeps earlier addresses working for the same owner. Previously claimed addresses remain reserved after account deletion to prevent reassignment; their account association is removed and the page becomes unavailable.
Encrypted backups
When independent backups are enabled, database recovery copies are encrypted before upload to a private Tigris object-storage bucket separate from the running Fly volume. They use a seven-day rolling retention target, with pruning after verified backup runs. Deleted records may remain in encrypted recovery copies until rotation; restoring a backup requires the separately held encryption key.
Providers and storage
The application and database run on Fly.io; sign-in uses Google and AI conversations use OpenAI. Reviewed introductions retrieved by your connected agent (Grok Bot or Muse) are also processed under that provider’s policies. Provider policies apply to their processing and retention. Revoking a connector prevents future retrieval but cannot remove copies an owner or their agent already received.
Read OpenAI’s data controls and Google’s privacy policy. Updated 26 October 2026.