privacy notice
this notice explains what scrubbr.ai collects when you connect gmail, google calendar, and slack to your scrubbr.ai account; how we use that data; and the rights you have over it.
1. who we are
scrubbr.ai is operated by scrubbr.ai llc, a delaware company. we are the controller of personal information you provide to use scrubbr.ai. you can reach us at privacy@scrubbr.ai.
2. information we collect
account information
name, email, profile photo, organization, and any preferences you set. we receive these from google when you sign in.
gmail content (with your permission)
once you authorize gmail through google oauth, scrubbr.ai reads message metadata (sender, recipient, subject, timestamps, labels) and message body content so the product can prioritize, summarize, and draft replies. attachments are never downloaded or stored.
slack content (with your permission)
once you authorize slack, scrubbr.ai reads message metadata and message body content from channels and direct messages you have access to. scrubbr.ai uses your personal user token, not a bot token; messages you send through scrubbr.ai appear in slack under your name. attachments are never downloaded or stored.
google calendar content (with your permission)
once you authorize google calendar, scrubbr.ai reads event metadata (titles, attendees, times, locations, descriptions) so the product can surface upcoming meetings, generate prep context, and reflect schedule conflicts in your feed. attachments linked to events are never downloaded.
google chat content (with your permission)
once you authorize google chat, scrubbr.ai lists the spaces and direct messages you belong to, reads message text in those spaces, and sends replies on your behalf when you approve a draft. attachment content in chat messages is never downloaded or stored. if you message someone you have never chatted with before, scrubbr.ai creates a new 1:1 direct message space on your behalf.
derived content
summaries, draft replies, priority tiers, and embeddings (numeric vectors used for retrieval) generated by AI from the content above. derived content is associated with your account and used only to power your scrubbr.ai experience.
operational data
logs, error reports, and product usage events. we do not include message content in logs.
communications with us
messages you send to support, sales, or any of our team email addresses.
2b. google oauth permissions
when you connect your google account, scrubbr.ai requests the following permissions. we request only what is necessary to deliver the features you use.
| permission | why scrubbr.ai needs it |
|---|---|
| gmail.modify | read, classify, and surface messages; archive and mark read on your behalf; store ai drafts in your gmail drafts folder; send approved replies as you. we do not permanently delete messages and we do not send without an explicit action from you. |
| calendar.events | read upcoming events for schedule view and pre-meeting briefings; create, update, and delete events when you schedule or cancel from inside scrubbr.ai. |
| chat.spaces.readonly | list the google chat spaces and dms you belong to so scrubbr.ai knows which conversations to sync. no message content is accessed through this permission. |
| chat.messages | read chat messages for the unified inbox; send approved draft replies as you. message text only — attachment content is never accessed. |
| chat.spaces.create | create a new 1:1 direct message space when you send a chat message to someone you have never chatted with before. only triggered by an explicit send action from you. |
| chat.memberships | read the membership (list of people) in chat spaces you already belong to. used to show the member roster of a chat space and power @mention autocomplete when you compose a chat reply. we do not list memberships of spaces you are not in, we do not modify any membership, and we do not persist the roster on our servers — results are rendered live in your browser only. |
| directory.readonly | read the names, email addresses, and basic organizational info (job title, department, photo) of users in your google workspace organization. used only to suggest your teammates when you click "view teammate calendar" so you can check their availability while scheduling. we do not access phone numbers, addresses, login activity, or any other extended profile data. results are rendered live in your browser and not stored on our servers. |
| admin.directory.resource.calendar.readonly | read the conference rooms (resources) configured in your google workspace organization, so we can suggest them when you pick a meeting location. we access only room metadata (name, email, capacity, features, building, floor) — no user data, no calendar event content. results are rendered live in your browser and not stored on our servers. |
| contacts.other.readonly | read your "other contacts" — people you have emailed but have not added to your saved google contacts. used only to display profile photos for the senders in your unified inbox, so an avatar appears next to each message instead of an initials placeholder. we read profile photo urls only. no image bytes ever touch scrubbr.ai's servers — your browser fetches the photo directly from google's cdn. we do not access phone numbers, addresses, or any other contact details. |
| gmail.settings.basic | read your gmail send identities (including the signatures you have configured for each), and your vacation responder status. used to append your real gmail signature to messages scrubbr.ai sends, to reply from the correct alias when a thread arrived at one of your alternate addresses (e.g. founder@), and to surface a "you are set to out of office until [date]" reminder on your dashboard. this permission is read-only — scrubbr.ai never modifies your gmail settings. |
| calendar.calendarlist.readonly | read the list of calendars you have access to in google calendar — your primary calendar, team calendars, shared client calendars, holiday calendars, and any conference-room or resource calendars. used to sync events from each calendar into scrubbr.ai's unified view and to power the "calendars" toggle list in settings, where you can mute noisy calendars. we do not modify calendar memberships. |
| calendar.settings.readonly | read your google calendar preferences — specifically your working hours and default timezone. used so that when scrubbr.ai proposes meeting slots ("find me time this week"), the proposals respect your actual working hours and timezone instead of guessing. read-only — scrubbr.ai never modifies your calendar settings. |
| chat.messages.reactions | read and add emoji reactions on google chat messages, so the reaction bar in scrubbr.ai's unified inbox works the same way it does for slack messages — you can tap an emoji to react, and reactions other people leave on chat messages are visible to you. reactions are scoped to messages in spaces you already belong to. we never modify the content of a chat message; only your reactions. |
you can review and revoke these permissions at any time via your google account permissions page or via scrubbr.ai settings → integrations.
2c. google api services user data policy
scrubbr.ai's use and transfer to any other app of information received from google apis will adhere to the google api services user data policy, including the limited use requirements. specifically: we use google data only to provide user-facing features in scrubbr.ai; we do not use it for advertising; we do not sell it; we do not allow humans to read it except as required for security or legal compliance.
3. how we use your information
- to operate scrubbr.ai: prioritization, drafting, calendar prep, search, and your settings.
- to improve the specific features we provide to you, based on your usage of those features. we do not train AI models on your content (see section 6).
- to communicate with you: service notices and your support requests. we do not send marketing without consent.
- to keep the service safe: abuse detection, integrity, and legal compliance.
4. how we share your information
sub-processors
we use the providers below to operate scrubbr.ai. each is contracted to handle data only on our instructions and for the purposes of running scrubbr.ai. the current list and what each does:
- amazon web services (us): application hosting, compute, queues.
- supabase (us): managed postgres database, authentication.
- openrouter (us): AI model gateway. routes requests to anthropic and openai.
- anthropic (us): large language model provider (claude family). processes message content to generate summaries, priority tiers, and drafts. does not retain content for training.
- openai (us): embedding model provider. processes message text to generate retrieval embeddings. does not retain content for training.
- stripe (us): billing and payments. card data goes directly to stripe; scrubbr.ai never sees it.
- google (gmail and calendar apis): your data source. governed by your existing google account.
- slack technologies (slack api): your data source. governed by your existing slack workspace.
the canonical list lives at scrubbr.ai/trust and is updated when sub-processors change.
legal disclosures
we will disclose information when required by valid legal process, after reviewing the request and notifying you where allowed.
business transfers
if scrubbr.ai is acquired or merges, your information may transfer as part of that transaction, subject to this notice.
5. how long we keep it
message and calendar content is retained on a rolling 90-day window for active accounts so the product can reason about recent context. derived summaries and embeddings are retained while your account is active.
if you cancel or delete your account, we delete your account data within 30 days, except where retention is required by law. you can request export or deletion at any time at privacy@scrubbr.ai.
6. AI training and your data
scrubbr.ai does not use your content to train any AI model, ours or any third party's. anthropic and openai are contractually prohibited from training on data sent through our enterprise integrations.
7. how we protect it
data is encrypted in transit (tls 1.2 or higher) and at rest (aes-256 with provider-managed keys). access to production systems is restricted to a small set of authorized personnel and is audited. our security policies are aligned with nist sp 800-171 rev. 2 and structured to support a soc 2 type II audit. attestation is in progress.
raw email and slack message content is invisible to scrubbr.ai employees at the application layer. employee tooling shows AI summaries and counts only.
8. where your data lives
scrubbr.ai stores data in the united states. AI processing is performed in the united states by anthropic and openai. if you are located outside the united states, your data will be transferred to and stored in the united states under appropriate safeguards.
9. your privacy rights
depending on where you live, you may have rights to access, correct, export, or delete your information, and to object to or restrict certain processing. contact privacy@scrubbr.ai to exercise any of these.
we do not sell your personal information. we do not share it for cross-context behavioral advertising.
10. cookies and tracking
scrubbr.ai uses a small number of strictly necessary cookies to keep you signed in and to remember preferences. we do not use third-party advertising trackers on the marketing site.
11. children
scrubbr.ai is not directed to children under 16. we do not knowingly collect data from anyone under 16.
12. changes to this notice
when we make material changes, we will update the effective date and notify active customers by email. previous versions are available on request.
13. contact
questions, requests, or complaints: privacy@scrubbr.ai. for security reports: security@scrubbr.ai. for general support: support@scrubbr.ai.