Privacy Policy
Local review draft — Last updated: October 20, 2026 (UTC) — not approved for publication
This Privacy Policy explains how VectorOS (“VectorOS,” “we,” “us”) collects, uses, stores, and protects your information when you use the VectorOS application and website. We aim to collect only what we need to run the product for you. This policy describes our data practices; it does not replace consent required by law.
Information we collect
Account and authentication. Your email address and the authentication credentials used to sign in. Passwords are handled by our authentication provider and stored in hashed form; we never receive or store your password in plain text.
Profile. Your display name and preferences such as time zone and daily-capacity settings.
Workspace content. The content you create or bring into VectorOS: opportunities, tasks and commitments, contacts and people, projects, goals, meetings, notes, and uploaded or imported files and documents. This is stored to provide the product to you.
Captures, commands, and their results. The text you type, paste, dictate or import into capture, brain dump and command surfaces, together with what VectorOS proposed in response, so you can review, confirm, or undo it later. This is processed on our own infrastructure — see Workspace intelligence and model tools below.
Usage and diagnostics. Limited operational information — such as feature usage, error and request logs, and approximate activity timestamps — used to operate, secure, and improve the service. VectorOS does not currently use third-party advertising or cross-site tracking.
Device and technical information. Standard information our infrastructure providers process to deliver and secure the service, such as IP address, device and browser type, and request metadata.
Referral measurement. If you reach VectorOS through one of our own referral links, we record that the link was used, and — if you go on to create an account — we link that record to your account so we can tell which listing brought you. The record contains no IP address, user-agent string or device fingerprint. Paid conversions are recorded only when Stripe confirms a real payment. This is described in full in our Cookie Policy.
Payment information. Subscriptions are processed by Stripe. Your payment card details are collected and stored by Stripe, not by VectorOS; we receive only limited billing status and identifiers needed to manage your subscription.
Imports and optional read-only connectors
VectorOS imports documents and files you upload, and text you paste or type. Google Calendar is available where enabled; Email (Gmail), Notion, Slack and CRM (HubSpot) are rolling out and require provider setup. With your authorization, Vector receives the source data permitted by the connection to generate reviewable proposals. We do not sell connected data or use it for advertising. No proposed work records are created until you review and confirm. Where Google Calendar connection is available, you can explicitly authorize a read-only, one-time snapshot of your primary calendar. Where configured, Gmail, Notion, Slack and HubSpot also offer a limited read-only initial snapshot with explicit review and confirmation.
Gmail reads up to 20 recent thread subjects, senders and snippets, without full bodies or attachments. Notion reads up to three recently edited shared pages and their first twenty top-level blocks. Slack reads up to fifteen recent messages from one joined public channel, excluding private conversations, files and replies. HubSpot reads up to twenty-five recently modified contacts and twenty-five deals; its permission can expose account-wide CRM data. Confirmed contacts are account-wide within your Vector account, consistent with the existing contact model; tasks and opportunities belong to your selected workspace. You must be authorized to import the selected source data. No external write permissions or background sync are used. Temporary access lasts at most ten minutes and no refresh credentials are retained. Reviews expire after fifteen minutes. Only confirmed sources are saved privately and linked to records; unchosen sources are not saved. Failed confirmation may retain a private source for reconciliation. Disconnect clears temporary local access; revoke the provider grant in the provider settings.
Calendar access reads the calendar name and identifier, and event identifiers, titles, descriptions, locations, dates, times, status, type and source links within your chosen window. No edit or send permission is requested. These fields produce deterministic proposals, not automatic work records. The connection is temporary (up to ten minutes), uses no refresh token, and does not enable background synchronization. Your review expires after fifteen minutes. Only explicitly confirmed work items and their source events are stored as a private linked snapshot; unchosen event content is not saved by this import. Failed confirmation can leave a private source file for reconciliation. Discarding a review does not delete a previously retained file or confirmed record.
Google data is used to provide this requested import and your confirmed workspace records, not advertising or AI-model training. Our use of information received from Google Workspace scopes will adhere to the Google User Data Policy, including its Limited Use requirements. You can disconnect the temporary connection in VectorOS and revoke Google authorization through your Google account. Disconnecting does not delete confirmed records or sources; deletion and export requests follow the processes below. See the Google Workspace user data policy.
Import and document reading can be incomplete or inaccurate — for example a missed, duplicated, or misread entity. VectorOS is designed so that you review imported information before it is saved as work records. Uploaded files and their metadata are stored before review and remain stored after discard; delete unwanted files from Import. You are responsible for confirming what you keep.
Founder Growth account connections
Where enabled for the founder, Growth can connect a business to its Google mailbox, Instagram professional account or LinkedIn member profile through official provider sign-in. These durable business connections are separate from the temporary, read-only customer imports described above. The founder chooses permissions and confirms the provider-observed account for the business. Provider consent alone does not approve a message or post.
Growth retains the selected account identifier, granted permissions, authorization expiry, connection history and encrypted server-side credentials needed to maintain authorized access. Gmail sending can be selected without mailbox reading. If reading is separately granted, Growth retrieves replies only from recorded campaign conversations; the provider permission can expose the mailbox more broadly. Recorded replies can include sender, subject, body, thread and message identifiers, dates and opt-out requests. They are retained privately with the business campaign to support reviewed replies, duplicate prevention and acknowledgment recovery.
Outbound email and social posts require approval of the exact content, account, destination and relevant time. Supported account metrics are retrieved only with the separately granted permission. Google mailbox data is used for the requested email features, not advertising targeting, sale of data or AI-model training. Incoming messages cannot authorize further work. Background work requires an available authorized worker and checks current business pause, connection status and approval before transmission.
Disconnect disables future local use and removes active access and refresh credentials. It does not erase recorded messages, campaign approvals or historical publication evidence; retention and deletion follow the processes below. A provider may also require you to remove its authorization in the provider account settings. Growth distinguishes local disconnect from provider-confirmed revocation.
Private connector activity history
Where enabled, Vector records the outcome, time and source-record count of connector reads and local access clearing, privately for your account and workspace. The last 50 events per provider are retained, without provider tokens, account names, source content or raw error bodies. Workspace reset and account/workspace deletion remove this metadata. It may be incomplete when recording was unavailable and does not establish current provider access or revocation.
Optional private meeting audio
Where enabled, you can import audio or record with your browser microphone for a private meeting you own in the current workspace. This does not join a call or record team meetings. Obtain every participant’s agreement before recording or requesting transcription. Browser microphone permission and your acknowledgement do not independently verify that everyone consented. The flow accepts supported audio up to 25 MiB and ten minutes.
Before upload, audio drafts are saved in this browser’s IndexedDB device storage, including captured chunks for recovery after interruption. Up to four device drafts are retained, each bounded to 25 MiB. Recovery depends on the same browser profile and available storage; export a copy before leaving. Renew the participant-consent acknowledgement after reload before uploading. Signing out, switching accounts or deleting server records does not automatically remove a device draft. Use Discard device draft to remove it; clearing browser data can also remove it. Protect a shared browser profile as you would other sensitive files.
Reviewing and confirming an upload reservation saves the selected source hash, byte size, format, duration, meeting/workspace identifiers, consent version and time, and retention choice. Upload is a separate step that sends the audio to private application Storage; queued transcription requires another explicit review and confirmation. Uploaded audio is not device-only. Its job status and timestamped transcript are stored in the application database under owner and workspace permissions. No tasks, sharing or publication are automatically authorized by recording, a transcript or a completed job.
Transcription is optional and requires a configured, separately operated local worker. That worker retrieves the uploaded audio through its configured loopback service, decodes it and runs the existing offline faster-whisper tiny.en model on the deployment machine. “Local” describes that processing runtime, which need not be your device. The worker creates temporary processing files and attempts to remove its job directory when it finishes; interruption can require operator cleanup. This path does not silently fall back to an external model service. An enabled upload or queued job does not prove that a worker is running or that transcription succeeded.
You choose 7 or 30 days of server retention when reviewing the request; that period starts when transcription is queued. An unfinished upload reservation expires after 15 minutes. Expiry makes a job due for cleanup, not proof its bytes have been erased: removal and finalization require the cleanup worker to run successfully. Cancelling processing does not delete the audio. An explicit deletion fences the job and requests removal of its fixed audio/transcript objects; it remains pending until the service verifies finalized deletion. Deleting a meeting or account must not be treated as proof that private Storage or device copies have already disappeared.
You can export the device audio draft and, when ready, the timestamped transcript. Downloads, backups and copies retained outside this workflow are not erased by deleting a server job or discarding the device draft. Review transcription accuracy before use; this feature does not provide verified speaker identification. Full account export and deletion requests still follow the request process below.
Separately reviewed notes, decisions and task proposals are stored as private evidence with their source transcript citations. They share the audio job’s access, retention and verified deletion lifecycle; saving a task proposal creates no task. Exported evidence is a separate copy that server deletion cannot remove.
Workspace intelligence and model tools
The customer workspace intelligence — importing and understanding your documents, extracting entities, recommendations and their explanations, signals, prioritization, brain dumps, and commands — is deterministic and runs on our own infrastructure. Your workspace data — your tasks, projects, people, notes and imported documents — is not sent to a third-party AI model, and it is not used to train anyone’s model. There is no AI model provider in this customer workspace planning and document-import path. Optional private meeting transcription uses a separate local speech model as described above.
Separate internal Growth tools may use configured search services and Anthropic for selected research and content inputs. Those administrative tools do not provide model processing for customer workspace imports or priorities. Their inputs and any resulting proposed actions still require review; provider availability is not proof of publication or a measured result.
Internal Growth media workers can use local narration and transcription models for selected production inputs. These local models are separate from customer workspace planning and from external model providers. Generated media requires separate review and publishing authority.
Optional customer response playback can send the selected response text to a configured local Kokoro worker to synthesize audio for an entitled workspace. This is separate from planning and imports. An unavailable local worker produces an explicit error, without silently using an external voice provider.
Browser voice dictation is a separate feature.Dictation into text fields uses your browser’s own speech-recognition capability — a browser feature, not VectorOS-hosted AI. The audio does not pass through a VectorOS server and we do not receive or store a recording. Depending on which browser you use, however, your speech may be sent to the browser provider and processed under its own terms and privacy policy. Typing instead of dictating avoids this processing. Meeting recording/import follows the separate private audio flow above. See our AI Transparency page.
How we use information
We use your information to operate and provide VectorOS — organizing your work and surfacing what deserves your attention; to process subscriptions and billing; to provide support; to maintain security, prevent abuse, and debug problems; to improve the service; and to comply with legal obligations. We do not sell your personal information, and we do not use your workspace content for advertising.
Legal bases for processing (EEA/UK)
Where the GDPR or UK GDPR applies, we process personal data on these bases: to perform our contract with you (providing the service and billing); our legitimate interests (securing, maintaining, and improving the service, and preventing abuse); your consent (for optional features you choose to enable, such as connected accounts or certain AI features); and compliance with legal obligations.
Service providers (processors)
The service is built to use the following providers. Their role depends on the features in use:
- Supabase — database, authentication, and file storage. Privacy.
- Vercel — application hosting and delivery. Privacy.
- Stripe — subscription and payment processing. Privacy.
- Anthropic and Tavily — optional research/content and search processing in separate internal Growth tools. Anthropic privacy; Tavily privacy.
Transactional and authentication emails (such as confirmations and password resets) are delivered through our authentication and infrastructure providers. If we add or change a material processor, we will update this list.
Retention and deletion
We retain your account and workspace data while your account is active. You can delete supported individual items in the app. Account deletion is handled by request, with identity verification and review of associated and shared records. Some records may need to be retained for legal, security, tax, or billing purposes (for example, transaction records Stripe is required to keep).
About backups. Copies of our database may be held by our infrastructure provider as part of its own operations, under its retention practices rather than a schedule we set. We do not publish a cycle time for those copies, because we would be quoting a number we cannot enforce. Deletion from our live systems is what we control and what an account-deletion request addresses, subject to applicable retention requirements.
Your rights and choices
Subject to applicable law, you may: access the personal information we hold about you; correct inaccurate profile information from within the app; delete individual items from within the app; export a copy of your workspace data; delete your entire account; and object to or restrict certain processing.
How to make these requests today. Correcting profile information and deleting individual items are done in the app. Account deletion and a full data export are handled by request rather than by a button: contact us using the address below. Applicable response deadlines and any lawful extensions depend on the request and jurisdiction. Deleting an account and cancelling a paid subscription are separate requests; use the billing controls or contact us about cancellation as well.
For any request, contact vectoros.app@gmail.com. We may need to verify your identity before acting on a request.
Regional privacy rights
Canada (including BC PIPA and PIPEDA where applicable). You may access and correct your personal information and ask about our handling of it. You may also contact the BC Office of the Information and Privacy Commissioner or the Office of the Privacy Commissioner of Canada, as applicable.
EEA / UK (GDPR). You have rights of access, rectification, erasure, restriction, portability, and objection, and the right to lodge a complaint with your local supervisory authority. Our legal bases are described above.
California (CCPA/CPRA). You have rights to know, access, correct, and delete personal information, and to opt out of “sale” or “sharing” of personal information. We do not sell or share your personal information as those terms are defined, and we will not discriminate against you for exercising your rights.
Security
We use technical and organizational safeguards designed to protect account data, including encryption in transit, authenticated access, and database access controls (row-level security) designed to scope records by owner, workspace membership and authorized sharing. Hosted configuration and operational safeguards require verification. No method of transmission or storage is completely secure, and we cannot guarantee absolute security. See our Security page.
International processing
Our providers may process and store information in Canada, the United States, and other regions where they operate. Transfer requirements depend on applicable law and provider arrangements. Contact us with questions about processing locations and safeguards. Your information may be processed in these locations.
Children
VectorOS is intended for adults and is not directed to children. Do not use VectorOS if you are under the age of majority in your jurisdiction, or under 16, whichever is greater.
Changes to this policy
We may update this policy from time to time. Material changes will be reflected by updating the date above. Where required, we will provide notice and obtain consent before a material change takes effect. Continued use does not replace consent or rights required by law.
Contact
Questions about this policy, or want to exercise a privacy right? Contact vectoros.app@gmail.com.
This page is part of the VectorOS Trust Center. Questions? Contact vectoros.app@gmail.com.