Privacy Policy

Premise AI | Outbounder is operated by Vincent Guihan. Send privacy questions, rights requests, or complaints to questions@vincentjlabs.com.

questions@vincentjlabs.com

A research run processes the brand or product, target account, buyer role or persona, sales method, deliverables, and target language.

Submitted information is used to research the seller and target account, retrieve and check public evidence, choose a relevant offer, and generate coordinated sales materials in the selected output language.

Research inputs are sent through the application server to the OpenAI Responses API, which may retrieve public web sources. The application uses stored background responses so it can poll for completion. OpenAI states that API data is not used to train models unless the account owner opts in. OpenAI documents a 30-day application-state retention period for stored Responses API requests and may retain abuse-monitoring logs for up to 30 days, or longer when legally required.

OpenAI’s API data-control documentation

OpenAI processes research requests. Cloudflare provides application infrastructure, database services, and human verification. These providers may process information outside the user’s province, state, or country under their own contractual and legal safeguards.

Cloudflare Turnstile Privacy Addendum

The selected interface language is stored in this browser. English is the source language; Premise AI translates non-English interface and support copy with OpenAI models. The selected kit language is included in the research request and saved with the kit.

The current browser stores kits, the current draft, theme, language, sidebar preferences, usefulness-response state, and private shared-link controls. Outbounder does not impose a fixed kit count; the browser’s local-storage quota is the practical limit. Existing kits remain unchanged if a new or copied kit cannot be saved. Kits do not synchronize.

Copy link creates a read-only snapshot in the application database and copies its unguessable public URL. Anyone with that URL can view the snapshot. The creating browser keeps a separate private revocation control so the current link can be revoked. Otherwise it expires after 30 days. Clearing browser data removes that local revocation control but does not revoke the server snapshot.

The database stores aggregate usage and explicit feedback counters, request identifiers, provider job identifiers, status, timestamps, temporary completed results, and user-created read-only snapshots. While a response is being validated or recovered, its request record may temporarily contain generated kit content and validation findings; a completed exact result is capped at 320 KB. The provider job identifier is the recovery access value, so anyone with the recovery URL can retrieve the result until cleanup. Request records older than seven days are removed during the next accepted generation, so removal does not necessarily occur at the exact seven-day mark. Shared-link abuse protection uses a daily SHA-256 hash of the requesting IP; the raw IP is not written to that database. Shared snapshots are separate from recovery results and are removed when revoked or expire after 30 days. Identifier-free aggregate counters may be retained for operational history and reporting.

After a saved kit receives a deliberate Yes response, an optional follow-up records one of four time-saved ranges: about 15 minutes, 15 to 60 minutes, 1 to 2 hours, or more than 2 hours. Yes or No and the optional range are sent in separate one-field requests and stored only as daily aggregate counters. Neither payload contains a kit ID, browser or session ID, form field, kit content, source URL, request ID, cookie, IP address, or free-form text. The current browser remembers both selections to reduce repeat submissions.

The public Usage page reads aggregate operational counters for completed generation, accepted starts, broad completion-time bands, deliberate Yes or No usefulness responses, and reported time-saved ranges. The helpful share and median reported time-saved band appear only after at least five relevant responses. The page reports events and responses, not users, customers, accounts, individual kits, or sales results, and adds no passive tracking.

View public usage proof

Cloudflare Turnstile receives a verification token and may receive the request IP to confirm human interaction. The public widget key is delivered at runtime; the private validation key remains server-side.

The application does not request payment information, address books, email or social credentials, customer contact lists, or an OpenAI API key from the seller. Do not submit confidential customer data or sensitive personal information.

Premise AI | Outbounder creates drafts. It does not send email, place calls, post social messages, or contact a target account.

Clearing the site’s browser data removes saved kits, the current draft, display preferences, feedback state, and private shared-link controls. Revoke a current shared link from Share before clearing browser data; otherwise the snapshot remains available until its 30-day expiry. Provider-side requests and operational logs remain subject to configured OpenAI and hosting-platform controls.

Depending on applicable law, a person may ask about access, correction, or deletion of personal information and may challenge these practices. Contact questions@vincentjlabs.com. Browser-local kits are controlled on the user’s device, and identifier-free aggregate counters cannot be connected to a browser, so the operator may not be able to identify those records as belonging to a requester.

Contact Vincent Guihan

Material changes will be posted on this page with a revised effective date. Continued use after a change is subject to the updated policy.