Privacy Policy

Welcome to Kabiri, operated by Kabiri Workspace Limited ("we", "us", "our"), a company registered in England and Wales (Company No. 17140110) with its registered office at Bartle House, Oxford Court, Manchester, England, M2 3WQ. We are committed to protecting your privacy. This Privacy Policy explains how Kabiri Workspace Limited, as the data controller, collects, uses, discloses, and safeguards your information when you visit our website at kabiri.co (the "Site") and use our services (the "Service"). Please read this privacy policy carefully. If you do not agree with the terms of this privacy policy, please do not access the site or use the service.

1. Information We Collect

We may collect information about you in a variety of ways. The information we may collect includes:

• Account Data: When you create an account, we collect your name, email address, company or organisation name, timezone preference, and other profile information you provide during registration.
• Derivative Data: Information our servers automatically collect when you access the Site, such as your IP address, browser type, operating system, access times, and the pages you have viewed directly before and after accessing the Site.
• Integration Data: When you connect third-party services (such as Gmail, Google Calendar, Google Drive, Google Tasks, iCloud Mail, Calendar, Reminders and Contacts, Dropbox, Slack, Canva, Todoist, Asana, Jira, Linear, ClickUp, or other supported integrations), we access and process data from those services on your behalf. This may include emails, calendar events, files, tasks (including the tasks you sync from Todoist, Asana, Jira, Linear or ClickUp), contacts, messages, and details of your designs. This data is used only to provide the features you use. It is never used for advertising and never sold. How we handle data from your Google account in particular is set out under Google user data in Section 2.
• App-Specific Passwords: Some services have no authorisation flow we can use and are connected with a password instead - iCloud (Mail, Calendar, Reminders and Contacts), and mailboxes connected over IMAP. For iCloud we ask for an app-specific password (a separate credential you generate at Apple and can revoke without changing your real password). For other IMAP mailboxes we recommend an app-specific password wherever your provider offers one, but the form also accepts your normal mailbox password, and if that is what you enter, that is what we store. It is encrypted with a key held in Google Cloud KMS, stored only in the UK / EU (europe-west2), never readable by your browser, and used only to reach that service on your behalf. Disconnecting deletes our copy immediately. We cannot revoke the password at the provider's end, so we recommend deleting it there too. For iCloud Calendar, Reminders and Contacts we store no events, reminders or contacts on our servers at all - each request fetches what it needs and passes it to your browser. iCloud Mail is handled like any other mailbox connected over IMAP.
• Server-Side Search & Sync Index: To keep your unified inbox in sync across your devices, we store email metadata (sender, recipients, subject, the short snippet of the message text that your provider supplies, timestamps, labels, and the headers that link replies together) on our servers, on every plan. On paid plans and during the free trial we additionally build a server-side search index so full-text search is fast and works on every device; this index holds search terms extracted from your messages and a short preview (up to roughly 300 characters) of each message body. This data is stored in our database in the UK / EU (europe-west2), used solely to provide search and sync, and kept until you delete the message, disconnect that mailbox (which deletes its copy straight away) or close your account. If you move from a paid plan to Free, the search index is removed shortly after that mailbox next syncs.
• Briefing: When you start a briefing on the dashboard, the Kabiri Assistant reads what it needs from your workspace at that moment and sends it to the Gemini API with the request: your recent email (the messages of up to eight of your most important conversations, including what you sent and to whom), your calendar, tasks and chat, and the alerts from your connected task apps (who mentioned you, commented or assigned you something, and where). It also reads the open action items in your Kabiri meeting notes for your next few meetings and the titles of files attached to them. If you turn on Read documents attached to meetings, it also reads up to three Google Docs attached to those meetings and passes on only their open action items. Encrypted notes are never read, and while note encryption is on, attached documents aren't either. Kabiri does not store the messages or documents the briefing reads, and the briefing it writes is kept on your device. When you open it in the Kabiri Assistant, it is also saved in that chat, like any other assistant message: in your Kabiri account (Firestore, europe-west2), until you delete the chat, auto-delete removes it, or you close your account. So that the next briefing doesn't repeat itself, Kabiri does remember what each briefing covered: the subjects and sender names of the conversations and the names of people it mentioned, the titles of tasks, and the titles and times of the day's meetings, never message text. That memory is kept on your device and in our database (europe-west2) for the last seven days briefed. If you turn on Scheduled briefing (paid plans, off by default), the same happens on our servers at the time you chose, without Kabiri open: the Kabiri Assistant reads the newest messages of up to eight of your most important Gmail conversations, your Google and Outlook calendars, your Kabiri events, tasks, chat and alerts from your task apps, and sends them to the Gemini API with the request. Those messages are read for that one request and are not stored; scheduled briefings don't read meeting notes or attached documents. The briefing it writes, with the subjects and senders it refers to, is stored in our database (europe-west2) for 7 days and then deleted automatically. Nothing is read while the Kabiri Assistant is switched off in your privacy settings.
• Notification tokens: If you turn on notifications on a device (the Briefing's Notify me when it's ready, or live Slack updates), we store a push token for that device with its browser type and when it was last used. Notifications are delivered through Google Firebase Cloud Messaging, which may process them outside the UK / EU. A "briefing is ready" notification carries counts only, never names, subjects or message text, and a file request's carries no names or file names; a Slack mention notification includes up to 140 characters of the message. A token is deleted when you sign out on that device, when the browser stops accepting it, or when you close your account.
• Usage Data: Information about how you use the Service, including features accessed, searches performed, and interactions with the Kabiri Assistant. If you accept analytics cookies, the app also keeps daily counts of which parts of Kabiri you open, how long you actively spend in each, how many emails you send or open and how many tasks, notes, boards, events and contacts you create, and how many of the Kabiri Assistant's drafts you send, open to edit or discard. These are counts and times only - never what you wrote, sent or opened, or who it involved. We use them to learn which features are useful, and our support staff may look at a 30-day summary for your account when helping you. They are deleted after 90 days, and turning analytics off deletes them straight away.
• Payment Data: If you subscribe to a paid plan, payment information is collected and processed by Stripe. We do not store your payment card details on our servers.
• Waitlist Data: If you sign up for our waitlist, we collect your email address and a timestamp of when you joined.
• Local Browser Storage: To make the Service fast and to let you read your inbox while offline, we cache a copy of your synced data - including email metadata and message bodies, calendar events, contacts, and attachments - in your browser's local storage (IndexedDB and localStorage) on each device you use. This cache lives only on your device and is not transmitted back to us. Clearing your browser's site data for kabiri.co (or signing out) removes it.

2. Use of Your Information

Having accurate information about you permits us to provide you with a smooth, efficient, and customised experience. Specifically, we may use information collected about you to:

• Provide, operate, and maintain the Service, including displaying your connected emails, calendars, files, and messages in the unified workspace.
• Power the Kabiri Assistant, which processes your queries and workspace data to provide smart suggestions, draft responses, and organise information on your behalf.
• Email you regarding your account, subscription, or waitlist status.
• Send periodic emails containing company news, updates, or related product information.
• Monitor and analyse usage and trends to improve your experience with the Service.
• Respond to enquiries and customer service requests.
• Comply with legal obligations.

Under UK GDPR we only process your personal data where we have a lawful basis to do so. The basis we rely on depends on the purpose:

• To provide and operate the Service (your account, connected emails, calendars, files, messages, and the Kabiri Assistant) - performance of a contract.
• To send marketing emails and process waitlist sign-ups - consent, which you can withdraw at any time.
• To secure the Service, prevent abuse, and analyse and improve usage - our legitimate interests in running and improving a safe, reliable Service, balanced against your rights.
• To meet our legal, tax, and accounting obligations - compliance with a legal obligation.

Email read receipts

Kabiri's email composer includes an optional read-receipt feature. It is off by default and only applies when the sender switches it on for an individual message. When enabled, the outgoing email contains a tiny invisible image; when the recipient's email app loads that image, we record that the message was opened.

• For each tracked message we store the recipients and subject (so the sender can see which email was opened), when it was sent, and the open count and the time of the most recent open. We do not store the recipient's IP address, device details, or location, and we do not share open data with anyone other than the sender. The record is kept until the sender closes their Kabiri account.
• If you receive a tracked email, your email app may block the image automatically (most webmail services load images through a proxy or let you turn image loading off), which prevents or masks the receipt.
• We process this data on the basis of our users' legitimate interests in knowing whether important mail arrived, balanced against recipients' rights by keeping the data collected to the bare minimum. To ask about or object to this processing, contact privacy@kabiri.co.

Company logos and website icons

When an email comes from a company, Kabiri can show that company's logo in place of the sender's initials. Company cards in Contacts show the icon of the company's website, and team bookmarks show the icon of the site they link to. To find these, Kabiri's servers (never your browser, and never a third-party icon service) look up the web address - for example stripe.com - and fetch the small icon the site publishes for itself.

• Only the web address (domain) leaves your account - never a full link, your email address, the sender's address, your messages or anything else about you. The request comes from our servers, so the company cannot see your IP address or tell who looked, or when.
• On emails, we only show a logo for companies whose email is protected against forgery (a published DMARC policy that rejects or quarantines forged mail), never for personal addresses such as Gmail or iCloud, never for your own colleagues, and never in Spam.
• Logos are kept on our servers for up to 30 days, filed under a one-way hash of the web address and shared by everyone who uses Kabiri. Nothing in that store says who received mail from whom.
• We do this on the basis of our legitimate interests in helping you recognise who an email is from at a glance. Because only a company's public web address is used, and nothing that identifies you, we do not offer a switch for it; you can still object to this processing at any time by writing to privacy@kabiri.co. Logos are the trademarks of their owners and are shown only to identify the sender.

File requests

A Kabiri user can make a file request: a link (kabiri.co/r/...) that anyone can open to send them files, with no Kabiri account needed. The files go into a folder the requester chose in their own Google Drive, Dropbox or OneDrive. If you send files through a request, you are the person this section is about, and the person who asked is "the requester".

• What we collect: the name and email address you give on the page (both are needed to send), the name, size and type of each file you send, when you sent it, and a one-way hash of your IP address. The hash is used to stop the page being abused; your actual IP address is not stored and cannot be recovered from it.
• Where the files go: straight from your browser to the requester's own storage account. Kabiri asks that account for a one-time upload address and hands it to your browser; the file itself never passes through Kabiri's servers and is never stored by Kabiri, and Kabiri does not open it. Your files land in a folder named after the name you gave (with a short code, so two people with the same name stay apart), inside the requester's folder. A file bigger than the request allows, or past its file limit, is moved to the requester's bin, where they can restore it. Once they are there, they are the requester's: what happens to them is up to the requester, under the terms they have with Google, Dropbox or Microsoft, which may keep them outside the UK / EU.
• Who sees your details: the requester, who sees your name, email address and the files you sent in Kabiri, and gets a notification saying how many files you sent and your name. The notification sent to their phone or computer carries no names or file names. If the requester asks the Kabiri Assistant about their file requests, what it needs to answer (such as your name and the file names) is sent to the Gemini API, as described in Section 3. Kabiri never uses your details for marketing, never adds you to a mailing list and never sells them.
• How long we keep it: the record of what was sent (your name, email address, the file details and the hashed IP address) is kept while the request exists, including after it closes, so the requester can see who sent what. It is deleted when the requester deletes the request, or with the requester's Kabiri account when they close it (within 30 days, or at once if they choose immediate erasure). Deleting it from Kabiri does not remove the files from the requester's storage.
• Why we do it: for the requester, file requests are part of the Service we provide them (performance of a contract). For you, we rely on legitimate interests: the requester's, in receiving the files they asked for and knowing who sent them, and ours, in keeping the page from being abused, and we collect only what those need. You can ask about, see, correct or object to what Kabiri holds about you at privacy@kabiri.co. To have the files themselves removed, ask the requester: they are in the requester's storage, and only the requester can decide to delete them.

Google user data

Kabiri'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. The use of information received from Google Workspace APIs (Gmail, Google Calendar, Google Drive and Google Tasks) will also adhere to the Google Workspace API User Data and Developer Policy, including the Limited Use requirements.

This section explains exactly what Kabiri does with data from your Google account. It applies when you sign in with Google, connect a Google account in Connected Apps, or connect Google Analytics to the dashboard. Where it is more specific than the rest of this policy, this section is the one that applies to your Google data.

• What you choose to connect: Signing in with Google gives us only your name, email address and profile picture. When you connect a Google account, you choose which of Gmail, Google Calendar, Google Drive and Google Tasks to connect, and Google shows you each permission before you approve it. Some features read your data in your browser, which talks to Google directly; others run on our servers, which use the access you granted while you are away. Both are described below.
• Gmail: Kabiri reads your messages so you can read, search and reply to them in Kabiri; organises them when you ask (labels, archive, trash, drafts); sends the email you write; and can import your Gmail signature. Your browser reads full messages directly from Gmail and keeps a copy on your device (see "Local Browser Storage"). Our servers keep the synced copy and search index described under "Server-Side Search & Sync Index", send the mail you write, send automatic replies you set up (holiday replies and VIP acknowledgements), check attachments for malware (the file is read for the check and not kept), archive, label and file the mail you approve in Organise your inbox (and, only if you switch it on, move suspected spam from the last 90 days to Gmail's own Spam folder, which Gmail empties after 30 days), and, if you turn on Scheduled briefing, read the newest messages of up to eight conversations at the time you chose. Kabiri never permanently deletes your email.
• Google Calendar: Kabiri reads your calendars and events to show them, and creates, changes, deletes and replies to events when you ask it to. Our servers also read your calendar: only your busy times (never titles or details) for your booking pages and for finding a time to meet; the events on calendars you own when a teammate you have shared your calendar with looks at it (see Section 3); the rest of today's events for a scheduled briefing; and, while Contact activity is on, your events from the past and next 30 days, to find your meetings with each contact. Our servers also add events to your calendar for you, such as confirmed bookings. Events are cached on your device. On our servers we keep a record of the events Kabiri added (the event's id, link and status), the titles and times of meetings in the Briefing's memory (the last seven days briefed) and, while Contact activity is on, the titles of your meetings with each contact.
• Google Drive: the Files tab is a full manager for your Drive, so Kabiri can list, open, upload, create, copy, share, move to trash and delete your files when you ask it to, and copy or move a file to another service you have connected, such as Dropbox or OneDrive. On paid plans, the names, types, dates, paths and links of the items in the folders you open (and your top-level Drive views) are kept on our servers so the Files tab loads quickly; a folder you have not opened for 30 days is removed. The Briefing looks for Drive comments that mention you and still need your reply. If you turn on Read documents attached to meetings, the Briefing reads up to three Google Docs attached to your next meetings in your browser and passes on only their open action items, along with the titles of files attached to the meeting. If you make a file request into Drive, our servers use your Drive access while you are away to check the folder you chose, make a folder for each person who sends files, open a one-time upload address for each file (the file goes from the sender's browser straight to your Drive) and check it arrived. File contents are not stored on our servers.
• Google Tasks: Kabiri keeps your Google Tasks and your Kabiri to-do list in step, both ways. Your tasks are stored in your Kabiri workspace in our database until you delete them or close your account.
• Google Analytics: if you add the Analytics widget, our servers read summary figures (such as visitors and page views) for the property you pick, with read-only access. We store only the connection and the property's id and name, not the figures.
• The Kabiri Assistant: some features send Google data to the Google Gemini API, our processor, to produce something for you: the Briefing, answers when you ask the Kabiri Assistant about your mail, calendar or files, translations and drafts you request, the sorting of senders when you run Organise your inbox (what kind of mail each sends, whether it looks like spam, whether it is work or personal, and folder suggestions, from sender names, addresses and a few subject lines, never message bodies), and, only if you turn them on (paid plans, off by default), priority labels for new mail, suggested replies and a daily inbox digest. Only what that feature needs is sent, and only for you. These background features stop while the Kabiri Assistant is switched off in your privacy settings. What they produce (priority labels, suggested replies and digests, which can include a subject or a short extract of a message) is kept in our database until you close your account. While an Organise run waits for you to approve it, the sender, subject and date of each message it would move are kept so you can see them on its review screen; they are deleted as soon as you approve or discard it. An Organise run's list of the messages it moved, and how it sorted each sender, are deleted after its 30-day undo window; a summary of each run (sender names and counts, and the addresses of the senders in any folder you chose) is kept until you close your account. Email rules it makes from your answers stay in your email settings until you delete them.
• Contact activity: on paid plans and during a trial it is on unless you switch it off (Contacts, More options, Relationship timelines; it stays off once you do). While it is on, Kabiri reads your last 90 days of email and keeps, with each of your contacts, the subjects and a short extract of your recent email with them and the titles of your meetings with them, up to 50 items per contact, until you close your account. Switching it off stops new items being added; what is already kept stays until you close your account. This is not sent to the Gemini API.
• No model training: Google user data is never used to develop, improve or train generalised or non-personalised artificial intelligence or machine learning models, whether ours, Google's or anyone else's. Kabiri does not train models at all. We use the Gemini API under Google's paid-service terms, under which Google does not use what we send to improve its products and keeps it only for a limited time to detect abuse and meet legal obligations.
• No advertising, no sale: we never sell Google user data, never use it to show you or anyone else adverts, never pass it to advertising platforms, data brokers or information resellers, and never use it to decide anyone's creditworthiness or for lending. The advertising tags described in Section 3 never run in the signed-in workspace and never receive Google user data.
• Who else receives it: we transfer Google user data only to provide the features described here (to Google Cloud, which hosts Kabiri; to the Gemini API; and, only if you turn on the grammar check, to LanguageTool, which checks the text you are writing, including a draft or reply that started in Gmail), to other people or services when you choose to (for example the email you send, a calendar you share, the busy times on your booking page, or a file you copy to another connected service), for security purposes such as investigating abuse, or to comply with the law. If our business is merged, acquired or sold, we will ask for your explicit consent before your Google data is transferred.
• People at Kabiri: no one at Kabiri reads your Google data, unless you first agree to it for specific messages, files or other data (for example when you ask our support team to look at a problem), it is needed for security purposes such as investigating abuse, it is needed to comply with the law, or it has been aggregated and anonymised for running the Service. Our support tools do not show the content of your mail, calendar or files.
• How it is protected: Google data travels encrypted between Google, our servers and your browser, and is encrypted at rest in our database in the UK / EU (europe-west2). The access Google gives us for your account (OAuth tokens) is additionally encrypted with a key held in Google Cloud KMS.
• Removing access and deleting your data: you can disconnect a Google account at any time in Connected Apps. That cancels Kabiri's access at Google and deletes our copy of its tokens, its synced mail metadata and its search index straight away; what the Kabiri Assistant produced from it and your Contact activity stay until you close your account. Removing the Analytics widget's connection deletes what we store for it. You can cancel Kabiri's access to any Google account, including Analytics, at myaccount.google.com/permissions. Closing your Kabiri account cancels access to the Google accounts you connected in Connected Apps and deletes all of your Google data, within 30 days or at once if you choose immediate erasure. To ask about your Google data, write to privacy@kabiri.co.

3. Disclosure of Your Information

We do not sell, trade, or otherwise transfer your personal information to third parties except as described in this Privacy Policy or with your consent. We may share information with:

• Service Providers: We use third-party service providers to perform functions on our behalf, including:
  • Google Cloud Platform / Firebase for data storage, authentication, and infrastructure, and Firebase Cloud Messaging for delivering the notifications you turn on.
  • Google Gemini API to power the Kabiri Assistant. Queries and relevant workspace context are sent to the Gemini API to generate responses, including the context a scheduled briefing reads on our servers at the time you chose (see "Briefing" above). Data sent to the Gemini API is processed in accordance with Google's data processing terms and is not used to train Google's models. While our application servers and database are pinned to europe-west2 (UK / EU), the Gemini API is a global Google endpoint and individual requests may be processed by Google infrastructure outside the UK / EU. Logs of assistant requests (used for rate-limiting and quota enforcement) are retained for up to 90 days; files you attach to an assistant turn are deleted from our storage within 30 days. Your Kabiri Assistant chats are stored in your Kabiri account (Firestore, europe-west2) so they follow you between devices, until you delete them, auto-delete removes them, or you close your account.
  • SendGrid for transactional email delivery.
  • LanguageTool to power the optional grammar & style check in the email composer, notes and sticky notes. When the feature is on, the text you are writing is sent to LanguageTool to be checked and a list of suggestions is returned. Notes you have encrypted are never sent. LanguageTool is EU-based (Germany) and does not retain text submitted through its API. The check can be turned off in Email Settings (for email) and in Settings › General › Writing (for notes and sticky notes); when it is off, nothing is sent. Spelling is checked by your browser's own spellchecker; Kabiri does not send it anywhere.
  • Stripe for subscription billing and payment processing. Payment card details are handled entirely by Stripe and are never stored on our servers.
  These providers only have access to the information necessary to perform their functions and are obligated to maintain confidentiality.
• Advertising Partners, only if you accept advertising cookies: We advertise Kabiri on Google and on Meta (Facebook and Instagram), and we load the measurement tags for both through Google Tag Manager. If - and only if - you turn on the Advertising category in our cookie banner, Google Ads and the Meta pixel receive a record that your browser reached a particular page on our site, together with the identifiers described in our Cookie Policy. This lets us tell whether an ad led to a sign-up, and lets those platforms show you Kabiri ads elsewhere. For the Meta pixel, Meta Platforms Ireland Limited is a joint controller with us for the collection and transmission of that data, under Meta's Controller Addendum; Google acts as an independent controller for Google Ads data under its controller-controller terms. Neither receives your account data, your files, or anything you store in Kabiri. If you decline, or withdraw later, no such data is sent at all - these tags never run on the signed-in workspace.
• Connected Services: When you connect third-party services (Gmail, Google Calendar, iCloud, Dropbox, Slack, Notion, Canva, Zoom, Todoist, Asana, Jira, Linear, ClickUp, etc.), data flows between Kabiri and those services via their official APIs. We only access the data necessary to provide the features you have enabled. You can disconnect any integration at any time in Connected Apps. For Google, see Google user data.
• Claude (optional dashboard widget): If you connect your own Anthropic API key to the Claude widget, the messages you send in that widget, together with the conversation so far, are sent from our servers to Anthropic to generate Claude's replies. This happens under your own Anthropic account and is governed by Anthropic's terms; usage is billed to that account, not by Kabiri. Your key is stored encrypted, is never sent back to your browser, and is deleted from Kabiri when you disconnect the widget. Your widget conversations are saved to your Kabiri account so they are available on your other devices. Anthropic's API is a global service, so requests may be processed outside the UK / EU.
• Canva (optional dashboard widget): Kabiri works with Canva through Canva's official Connect API. If you connect your Canva account, Kabiri reads the titles, thumbnails, last-edited dates and links of your Canva designs, and your Canva display name, so the widget can show your recent designs and open them in Canva. Kabiri asks Canva only for permission to read those details; it cannot create, change or delete anything in your Canva account. Your designs and their thumbnails are fetched from Canva each time the widget shows them and are not stored by Kabiri. What we do store is the connection itself: the access and refresh tokens Canva issues, encrypted and never sent back to your browser, and your Canva display name, so the widget can show which account is connected. Anything you type into the widget's search box is sent to Canva to find matching designs. When you open a design, you go to Canva, where Canva's own terms and privacy policy apply. Canva Pty Ltd is based in Australia and runs a global service, so this data may be processed outside the UK / EU. When you disconnect Canva from Connected apps, we revoke Kabiri's access at Canva and delete the stored tokens straight away. When you close your Kabiri account, we revoke that access at once, and the tokens are deleted with the rest of your account data. You can also remove Kabiri at any time from the connected apps in your Canva account settings. We never use your Canva data for advertising, sell it, or share it with anyone else. Canva is a trademark of Canva Pty Ltd. Kabiri is not endorsed, sponsored or otherwise approved by Canva.
• Instagram and Facebook (optional insights): Kabiri works with Instagram and Facebook through Meta's official Graph API, with one Facebook login. If you connect them, Kabiri reads, from Meta Platforms Ireland Ltd, the Facebook Pages you manage and chose on Facebook's consent screen, and the Instagram Business or Creator account linked to the Page you pick: each Page's name, follower and like counts, its views and post engagements over the last 28 days, its recent posts (text, picture, link and their likes, comments and shares), and the Instagram account's username, follower, following and post counts and its recent posts (caption, picture, link, likes and comments). This is so the optional dashboard widgets can show them and, on Premium, so the Kabiri Assistant can answer your questions about them. Kabiri asks Meta only for permission to read: it never posts, comments, replies or sends messages on Instagram or Facebook. Figures and posts are fetched from Meta each time they are shown or asked about and are not stored by Kabiri. What we do store is the connection itself: the access tokens Meta issues (encrypted, and never sent back to your browser), which last about 60 days, your name on Facebook, your Pages' names and your Instagram username, so we can show which accounts are connected; and a count of how many times the Kabiri Assistant has read them this hour. When you disconnect from Connected apps, we remove Kabiri's access at Meta and delete the connection straight away. When you close your Kabiri account, we remove that access at once, and the connection is deleted with the rest of your account data. If you remove Kabiri in your Facebook settings and ask Meta for your data to be deleted, Meta tells us and we delete the connection. Meta handles your accounts under its own terms and privacy policy, and its services run in the EU and the US, so this data may be processed outside the UK. We never use your Instagram or Facebook data for advertising, sell it, or share it with anyone else. This is separate from the advertising cookies described below, which you can decline. Instagram and Facebook are trademarks of Meta Platforms, Inc. Kabiri is not endorsed, sponsored or otherwise approved by Meta.
• Zoom (optional video calls for calendar events): Kabiri works with Zoom through Zoom's official API, so you can add a Zoom call to an event you create in Kabiri's calendar, and so a booking made through your booking page can come with one. Nothing reaches Zoom until you connect your Zoom account; when you do, Zoom asks you to let Kabiri create meetings on your account and read your basic Zoom profile. When you pick Zoom as an event's video call, or someone books a slot on a booking link of yours set to Zoom, Kabiri's servers send Zoom the event's title (as the meeting's topic), its description (as the agenda, if it has one) and its time, and Zoom creates the meeting on your account. Kabiri puts the meeting's join link on the event (and in its location if you left that empty), where everyone invited can see it. Kabiri never keeps or shows the meeting's start link, which would let anyone holding it run the meeting as you. What we store is the connection itself: the access and refresh tokens Zoom issues, encrypted and never sent back to your browser, and your Zoom user id, account id, name, email address and personal meeting id, so Connected apps can show which account is connected. Kabiri keeps no record of your meetings beyond the join link on each event, and never receives your recordings, chats or who attended. Zoom Communications, Inc. is based in the United States, so the meeting details above are processed there, under Zoom's own terms and privacy statement, unless your Zoom account keeps its data in another region. When you disconnect Zoom in Connected apps, we revoke Kabiri's access at Zoom and delete the stored tokens straight away; join links already on your events stay there. When you close your Kabiri account, we revoke that access at once, and the tokens are deleted with the rest of your account data. You can also remove Kabiri at any time from the apps you've added in your Zoom account. We never use your Zoom data for advertising, sell it, use it to train any model, or share it with anyone else. Zoom is a trademark of Zoom Communications, Inc. Kabiri is not endorsed, sponsored or otherwise approved by Zoom.
• Notion (optional, every plan): Kabiri works with Notion through Notion's official API. When you connect your own Notion workspace, you choose on Notion's consent screen exactly which pages Kabiri may see. Kabiri then reads those pages' titles, their content when you search them, ask the Kabiri Assistant about them or bring them into Notes, their last-edited dates and links, and your workspace's name and icon. Pages are fetched from Notion each time and are not stored by Kabiri, except the notes you choose to bring in, which become ordinary Kabiri notes in your account. What we do store is the connection itself: the tokens Notion issues, encrypted and never sent back to your browser, and your workspace's name and icon, so Connected apps can show which workspace is connected. Anything you type into Notion search in Kabiri is sent to Notion to find matching pages. When you ask the Kabiri Assistant about a Notion page, the text it reads is sent to the Gemini API with your request, as described in Section 3. Kabiri creates a page in your Notion workspace only when you choose Save to Notion (a Premium feature) and press Save: it writes that page's title and content (the note, email or event details you are saving, or a draft the Kabiri Assistant showed you) under the Notion page or database you pick, and remembers that place as your default if you ask it to. It does not change or delete anything else in your Notion workspace, except the rows of databases you choose to sync, below. On Premium, if you choose to keep imported notes updated from Notion, Kabiri checks which of those pages changed when you open Notes (at most every 15 minutes) or press Refresh now, and reads just those pages again to update your notes; it stops for a note once you edit it, or when you turn it off. On Premium, if you choose Contacts from Notion in Contacts, Kabiri reads the columns you match (such as name, email and phone) from the one Notion database you pick and adds those people to your contacts, one way; nothing is written back to Notion, and stopping it keeps the contacts already added. On Premium, if you choose to sync a Notion database with a Kabiri project (Connected apps > Notion > Sync a database), Kabiri reads that database's rows and their properties, with the names of the people on them (and their email addresses where Notion shares them with apps, kept on Kabiri's servers only to tell whether someone is you or a Kabiri teammate), keeps a copy up to date in your Kabiri project, and makes the changes you and your team make to those tasks in Kabiri in Notion too: creating, editing, ticking off and deleting rows (a deleted row goes to Notion's trash). It never reads or writes the page content inside a row. Stopping a database's sync keeps its tasks in Kabiri unless you choose to remove them, and disconnecting Notion keeps them. Notion Labs, Inc. is based in the United States, so this data may be processed outside the UK / EU. When you disconnect Notion from Connected apps, we delete the stored tokens straight away; notes you brought in stay in Kabiri until you delete them. When you close your Kabiri account, the tokens are deleted with the rest of your account data. You can also remove Kabiri at any time from the connections in your Notion settings. Your Notion data is never used to train, fine-tune or evaluate any model, never used for advertising, never sold and never shared with anyone else. A team's wiki can be connected to a Notion workspace separately, by the team's owner or an admin, once Teams opens. Notion is a trademark of Notion Labs, Inc. Kabiri isn't affiliated with or endorsed by Notion.
• Todoist (optional two-way task sync): Kabiri works with Todoist through Todoist's official API, to keep the Todoist projects you choose in step with Kabiri projects, both ways. When you make a new Kabiri project and ask for it, Kabiri also creates a Todoist project with the same name, and syncs it. When you connect your Todoist account, Kabiri asks Todoist for permission to read and change your tasks and projects, and to delete tasks. It stores the connection itself: the access and refresh tokens Todoist issues, encrypted and never sent back to your browser, and your Todoist account id, name and email address, so Connected apps can show which account is connected. What our servers receive. Todoist's access covers your whole Todoist account, not only the projects you sync, so our servers receive task data from across that account: the list of your projects and their sections when you choose what to sync, the lists of changed tasks we ask Todoist for to keep your synced projects up to date, and the notices Todoist sends whenever something changes anywhere in the account. We process that data only in passing, to find the changes to the projects you sync or, when you ask, for the Kabiri Assistant's search (below), and keep nothing from the projects you don't sync apart from the alerts below; from Todoist's change notices we look only at which project, task or comment changed. For a synced task whose comments changed, and for one someone opens in Kabiri, our servers then ask Todoist for that task's comments. Alerts from Todoist. Unless you switch them off in Settings > Notifications, our servers check your Todoist notifications (the ones Todoist itself shows you) about every 5 minutes for anything that mentions you, is assigned to you, or is a new comment on something of yours, including outside the projects you sync. For each one, Kabiri keeps a short alert for 30 days: the name of the person who did it, what kind it was, the task's key, title and place, and its link into Todoist. A comment is looked at only to see whether it names you. Kabiri never keeps a comment's text for an alert, and never writes anything to Todoist for one. Searching Todoist with the Kabiri Assistant. When you ask the Kabiri Assistant to find something in Todoist, our servers search your open Todoist tasks across every project, not only the ones you sync, and read back the matching tasks' titles, project names, priorities and due dates; for one task you ask about, also an extract of its description and its three newest comments. Never an email address, and nothing is written to Todoist. What comes back is sent to the Google Gemini API, our processor, for that one reply, as with any question you ask the Kabiri Assistant. Our servers don't store the search results: they are used for that reply, which you see in the chat like any other answer, and the only thing stored is a count of how many Todoist searches you made in the hour. What we keep. For each Todoist project you sync, Kabiri keeps its tasks in the Kabiri project you chose: titles, descriptions (with their formatting), due dates (repeating ones included) and their time zones, deadlines, durations, priorities, labels, sections, sub-tasks, whether each is done, who each is assigned to (by name), and their links into Todoist, along with the project's name and colour, its sections, and the colours of the labels it uses. People in the project. To show who a task is assigned to, our servers read the names and email addresses of the people your Todoist projects are shared with, and keep only those of the synced project. Their names are kept on the tasks. Their email addresses are kept on our servers only, with their Todoist ids, never shown in the app or sent anywhere, and used for one thing: to tell whether that person is you or, on a Kabiri team project, one of your teammates, by the email address their Kabiri account has verified. Nobody else is looked up. That record on our servers is deleted when you stop syncing the project, disconnect Todoist or close your account; tasks you keep still show the names. These tasks are stored with the rest of your workspace in the UK / EU (europe-west2), with a small record of what was last synced, so that a change on either side reaches the other. Changes you make to those tasks in Kabiri (adding, editing, completing or deleting them, and who they are assigned to) are sent to Todoist. Comments. The comments on synced tasks are kept in step both ways too. Comments written in Todoist, including by people who don't use Kabiri, are kept with the task in Kabiri: their text (with its formatting), when each was written or last edited, and the display name of whoever wrote it, never their email address. Comments written on a synced task in Kabiri are posted to Todoist from your Todoist account; one a teammate writes starts "From" and their name, so Todoist shows who said it. You can edit or delete your own comments in Kabiri and Todoist follows, and edits and deletions made in Todoist show in Kabiri; other people's comments can't be changed in Kabiri. Files and links attached to the task's comments in Todoist are shown by name, with a link that opens them in Todoist when Todoist itself stores them; nothing is copied into Kabiri's storage, and files you attach in Kabiri are not sent to Todoist. Only people who can see the Kabiri project see its comments and attachments. They are deleted with their task, and when you stop syncing a project or disconnect Todoist and choose to remove its tasks; tasks you keep keep their comments as ordinary Kabiri comments, with their authors' names, and their attachments as links into Todoist. If the Todoist project is shared, the tasks your Todoist collaborators add to it come into Kabiri too. If you sync a project in a Kabiri team workspace, everyone who can see that project sees its Todoist tasks, and the tasks your teammates add to it are copied into your Todoist account. Once in Kabiri, synced tasks are part of your workspace like any other task, so the Kabiri Assistant can see them when you ask it about your work. Stopping. When you stop syncing a project, its tasks stay in Kabiri as ordinary tasks unless you choose to remove them, and nothing is deleted in Todoist. When you disconnect Todoist from Connected apps, we stop syncing every project, revoke Kabiri's access at Todoist, and delete the stored tokens and sync records straight away. When you close your Kabiri account, we revoke that access at once, and the rest is deleted with your account data. You can also remove Kabiri at any time in Todoist under Settings › Integrations. When you open a task in Todoist, Todoist's own terms and privacy policy apply. Todoist Inc., the US company that makes Todoist, stores Todoist data in the United States, so data you sync is processed outside the UK / EU. We never use your Todoist data for advertising, sell it, or share it with anyone else. Todoist is a trademark of Todoist Inc. Kabiri is not created by, affiliated with, or supported by Todoist.
• Asana (optional two-way task sync): Kabiri works with Asana through the official Asana API, to keep the Asana projects you choose in step with Kabiri projects, both ways. When you make a new Kabiri project and ask for it, Kabiri also creates an Asana project with the same name, in the workspace and team you pick, and syncs it. Nothing syncs until you connect your Asana account and agree, on the connect screen, to Kabiri copying the tasks in the projects you pick, with their comments and the names and email addresses of the people on them, and keeping them in step both ways, including sending your changes and deletions in Kabiri to Asana. When you connect, Asana asks you to give Kabiri full access to your Asana account, because Asana has no narrower permission that covers everything a two-way sync needs; Kabiri uses that access only for the projects you sync. It stores the connection itself: the access and refresh tokens Asana issues, encrypted and never sent back to your browser, and your Asana user id, name and email address, so Connected apps can show which account is connected. What our servers receive. To let you choose what to sync, our servers ask Asana for the workspaces you belong to and the projects in them (their names and ids). For each project you sync, they ask Asana for its sections, its tasks (open ones, and those finished in the last 30 days) and their sub-tasks, its custom fields (their names, types and options), the project's colour, and the names and colours of the tags in its workspace that your labels use. For each task they also receive its formatted notes, who it is assigned to and who follows it (their names and email addresses, and the same for people named in its custom fields), the tasks it depends on, whether it is a milestone or awaits approval, and the time tracked on it. They set up a webhook for that project, so that Asana tells our servers when something in it changes, and check the project's change feed at Asana every few hours in case a notice went missing. From those notices we look only at which task, section or comment changed, then ask Asana for the latest version. For a task whose comments changed, and for one someone opens in Kabiri, they also ask Asana for the task's activity, of which we keep only the comments (with their authors' names and email addresses), and for the files and links attached to it. We keep nothing from the projects you don't sync, apart from the alerts below. Alerts from Asana. Unless you switch them off in Settings > Notifications, our servers check the tasks assigned to you and their new comments about every 5 minutes for anything that mentions you, is assigned to you, or is a new comment on something of yours, including outside the projects you sync. For each one, Kabiri keeps a short alert for 30 days: the name of the person who did it, what kind it was, the task's key, title and place, and its link into Asana. A comment's text is read only to see whether it mentions you. Kabiri never keeps a comment's text for an alert, and never writes anything to Asana for one. What we keep. For each Asana project you sync, Kabiri keeps its tasks in the Kabiri project you chose: titles, descriptions (with their formatting; anything Kabiri can't show, such as an image, is kept on our servers so it isn't lost when you edit the notes), due and start dates and times, sections, sub-tasks (two levels), tags (as labels, with their colours), priorities (when the Asana project has a Priority field), custom field values (as Kabiri fields), who each task is assigned to and who follows it (by name), the tasks each depends on, milestones, approval status, time tracked, whether each is done, and their links into Asana, along with the project's name and colour, its sections, its custom fields, its workspace id and up to 200 of the workspace's tag names, so that labels match. People in the project. Our servers keep the names and email addresses of the people named on the synced tasks (assignees, followers, people in custom fields, the authors of comments), with their Asana ids, and nowhere else: the tasks hold their names, and the email addresses are never shown in the app or sent anywhere. They are used for one thing: to tell whether that person is you or, on a Kabiri team project, one of your teammates, by the email address their Kabiri account has verified. Nobody else is looked up. That record on our servers is deleted when you stop syncing the project, disconnect Asana or close your account; tasks you keep still show the names. These tasks are stored with the rest of your workspace in the UK / EU (europe-west2), with a small record of what was last synced and the project's webhook secret (encrypted), so that a change on either side reaches the other. Changes you make to those tasks in Kabiri (adding, editing, completing or deleting them, their custom field values, who they are assigned to, who follows them and what they depend on) are sent to Asana. Deleting a task in Kabiri that is also in other Asana projects only takes it out of the synced one. Kabiri marks each task it creates in Asana with a Kabiri reference that only Kabiri can see, and adds a tag in Asana when you give a synced task a label Asana doesn't have yet. Repeats are not synced. Comments. The comments on synced tasks are kept in step both ways too. Comments written in Asana, including by people who don't use Kabiri, are kept with the task in Kabiri: their text (with its formatting), when each was written or last edited, and the display name of whoever wrote it, never their email address. Comments written on a synced task in Kabiri are posted to Asana from your Asana account; one a teammate writes starts "From" and their name, so Asana shows who said it. You can edit or delete your own comments in Kabiri and Asana follows, and edits and deletions made in Asana show in Kabiri; other people's comments can't be changed in Kabiri. Files and links attached to a task or its comments in Asana are shown by name, with a link that opens them in Asana; Asana's short-lived download links are never kept, nothing is copied into Kabiri's storage, and files you attach in Kabiri are not sent to Asana. Only people who can see the Kabiri project see its comments and attachments. They are deleted with their task, and when you stop syncing a project or disconnect Asana and choose to remove its tasks; tasks you keep keep their comments as ordinary Kabiri comments, with their authors' names, and their attachments as links into Asana. If the Asana project is shared, the tasks your Asana collaborators add to it come into Kabiri too. If you sync a project in a Kabiri team workspace, everyone who can see that project sees its Asana tasks, and the tasks your teammates add to it are copied into the Asana project. Once in Kabiri, synced tasks are part of your workspace like any other task, so the Kabiri Assistant can see them when you ask it about your work. Your Asana data is never used to train, fine-tune, test or evaluate any model, ours or anyone else's. Stopping. When you stop syncing a project, Kabiri removes its webhook at Asana, and its tasks stay in Kabiri as ordinary tasks unless you choose to remove them, with their custom fields as ordinary Kabiri fields, values kept, and their comments; nothing is deleted in Asana. When you disconnect Asana from Connected apps, we stop syncing every project, remove their webhooks, revoke Kabiri's access at Asana, and delete the stored tokens and sync records straight away. When you close your Kabiri account, we revoke that access at once, and the rest is deleted with your account data. You can also remove Kabiri at any time in Asana, under your profile photo › Settings › Apps. Who is responsible. Kabiri is the controller of the copy of your Asana data held in Kabiri, independently of Asana, Inc. This policy covers how Kabiri handles it; the Asana Privacy Statement does not. Asana, Inc. is not responsible for the privacy, security or integrity of your Asana data once it is outside Asana, in Kabiri. If we ever have a security breach that touches data from Asana, we will tell Asana, Inc. as well as you. Inside Asana, the Asana terms and Privacy Statement apply. Asana, Inc. (633 Folsom Street, Suite 100, San Francisco, CA 94107-3600, United States; privacy@asana.com, and its Data Protection Officer at dpo@asana.com) stores Asana data in the United States unless your organisation has chosen another region, so data you sync is processed outside the UK / EU. For transfers from the UK, Asana, Inc. relies on the EU-US Data Privacy Framework and its UK Extension, and on Standard Contractual Clauses with the UK Addendum where those don't apply. We never use your Asana data for advertising, sell it, or share it with anyone else. Asana, the Asana logo, and combinations thereof are trademarks of Asana, Inc. Kabiri is not affiliated with, endorsed or sponsored by Asana.
• Jira (optional two-way task sync): Kabiri works with Jira Cloud (sites at atlassian.net) through Atlassian's official Jira REST API, to keep the Jira projects you choose in step with Kabiri projects, both ways. If you are an administrator of a Jira site and ask for it when you make a new Kabiri project, Kabiri also creates a Jira project with the same name on that site, with you as its lead, and syncs it; for this, the connect screen also asks you to let Kabiri manage Jira projects. Nothing syncs until you connect your Atlassian account and agree, on the connect screen, to Kabiri reading the issues in the Jira projects you pick, with their comments and the names of the people on them (and their email addresses, where their Atlassian profiles show them), keeping a copy up to date in Kabiri, and making the changes you and your team make in Kabiri in Jira too: creating, editing, moving, commenting on and deleting those issues. Kabiri writes to Jira only the changes made in Kabiri to the projects you sync, and only as you: never to other projects, and never on its own initiative. When you connect, Atlassian asks you to let Kabiri read and change the Jira issues you can see, read the names of Jira users, manage webhooks (so that Jira can tell Kabiri when an issue changes) and stay connected while you're away; Kabiri uses that access only for the projects you sync and, when you ask, for the Kabiri Assistant's Jira search (below). A Jira connection made before 2 November 2026, when Jira synced one way only, stays read-only: Kabiri keeps reading from Jira but sends it nothing until you reconnect your Atlassian account and agree to the two-way sync on the connect screen. Kabiri stores the connection itself: the access and refresh tokens Atlassian issues, encrypted and never sent back to your browser, your Atlassian account id and name, so Connected apps can show which account is connected, and the Jira sites you can reach (their ids, addresses and names). What our servers receive. To let you choose what to sync, our servers ask Atlassian for the Jira sites you can reach and the projects on each (their names, keys and ids). For each project you sync, they ask Jira for its statuses (its workflow), issue types, priorities, fields (their names, types and options) and issue link types, and for its issues: each issue's summary, description, status, due and start dates, priority, labels, story points, components, fix versions and other fields, who it is assigned to, who watches it and who reported it (their names and Atlassian account ids, and their email addresses where their Atlassian profiles show them), its issue type, sprint, resolution and parent, the issues it links to, its comments (with their authors' names and account ids) and the names and links of its attachments. They set up one webhook for each Jira site you sync, covering only the synced projects on it, so that Jira tells our servers when an issue or a comment in them changes, and check each synced project every few hours in case a notice went missing. From those notices we look only at which issue or comment changed, then ask Jira for the latest version. We keep nothing from the projects you don't sync, apart from the alerts below. Alerts from Jira. Unless you switch them off in Settings > Notifications, our servers check the issues assigned to you, that you reported or that you watch, and their new comments and changes about every 5 minutes for anything that mentions you, is assigned to you, or is a new comment on something of yours, including outside the projects you sync. For each one, Kabiri keeps a short alert for 30 days: the name of the person who did it, what kind it was, the issue's key, title and place, and its link into Jira. A comment's text is read only to see whether it mentions you. Kabiri never keeps a comment's text for an alert, and never writes anything to Jira for one. Searching Jira with the Kabiri Assistant. When you ask the Kabiri Assistant to find something in Jira, our servers search every Jira site and project your connection can see, not only the ones you sync, and read back the matching issues' keys, titles, statuses, the names of the people they are assigned to, their priorities, types and when they last changed; for one issue you ask about, also an extract of its description and its three newest comments with their authors' names. Never an email address, and nothing is written to Jira. What comes back is sent to the Google Gemini API, our processor, for that one reply, as with any question you ask the Kabiri Assistant. Our servers don't store the search results: they are used for that reply, which you see in the chat like any other answer, and the only thing stored is a count of how many Jira searches you made in the hour, to keep within Atlassian's limits. What we keep. For each Jira project you sync, Kabiri keeps its issues as tasks in the Kabiri project you chose: their titles, descriptions (with their formatting; anything Kabiri can't show, such as an image or a panel, is kept on our servers so it isn't lost when you edit the description), statuses (the project's columns are the Jira project's own statuses), due and start dates, priorities, labels, field values (as Kabiri fields), who each is assigned to, watches and reported it (by name), the issues each links to, whether each is done, and their links into Jira, along with the project's name, statuses and fields. People in the project. Our servers keep the names and Atlassian account ids of the people named on the synced issues (assignees, watchers, reporters, people in fields, the authors of comments), and their email addresses when their Atlassian profiles show them (Atlassian shares an email address only where that person's profile settings allow). The tasks hold their names; the email addresses are kept on our servers only, never shown in the app or sent anywhere. They are used for one thing: to tell whether that person is you or, on a Kabiri team project, one of your teammates, by the email address their Kabiri account has verified. Nobody else is looked up. That record on our servers is deleted when you stop syncing the project, disconnect Jira or close your account, and a person's entry is erased when Atlassian reports their Atlassian account closed. When a project stops syncing or Jira is disconnected, the tasks Kabiri keeps no longer hold Jira people's names or ids: the assignees and watchers from Jira are taken off them, and comments by Jira people, and mentions of them, show as "Jira user". Telling Atlassian. Atlassian requires apps that store people's personal data to report it, so that it can tell them when someone's details change or their Atlassian account is closed. Every 7 days (or as often as Atlassian asks), our servers send Atlassian the Atlassian account ids Kabiri stores for your synced projects, and nothing else about those people. When Atlassian answers that an account is closed, Kabiri erases that person's name and email address everywhere it copied them from Jira: their comments then show as written by "Jira user", and they are removed from our record of the project's people. When it answers that their details changed, Kabiri reads them again at the next sync. These tasks are stored with the rest of your workspace in the UK / EU (europe-west2), with a small record of what was last synced, so that a change on either side reaches the other. Changes you make to those tasks in Kabiri (adding, editing, completing or deleting them, moving them between the Jira project's statuses, their field values, who they are assigned to, who watches them and what they link to) are sent to Jira, as your Jira permissions and the project's workflow allow. A move the workflow in Jira refuses is not sent: the task goes back to the status it has in Jira, and for a day Kabiri shows that it couldn't move it, with a link to the issue in Jira. Kabiri marks each issue it creates in Jira with a Kabiri reference that only Kabiri reads. Each task links to its issue in Jira. Repeats are not synced. Comments. The comments on synced issues are kept in step both ways too. Comments written in Jira, including by people who don't use Kabiri, are kept with the task in Kabiri: their text (with its formatting), when each was written or last edited, and the display name of whoever wrote it. Comments written on a synced task in Kabiri are posted to Jira from your Atlassian account; one a teammate writes starts "From" and their name, so Jira shows who said it. You can edit or delete your own comments in Kabiri and Jira follows, and edits and deletions made in Jira show in Kabiri; other people's comments can't be changed in Kabiri. Files attached to an issue in Jira are shown by name, with a link that opens them on your Jira site; nothing is copied into Kabiri's storage, and files you attach in Kabiri are not sent to Jira. Only people who can see the Kabiri project see its comments and attachments. They are deleted with their task, and when you stop syncing a project or disconnect Jira and choose to remove its tasks; tasks you keep keep their comments as ordinary Kabiri comments, shown as written by "Jira user", and their attachments as links into Jira. If you sync a project in a Kabiri team workspace, everyone who can see that project sees its Jira tasks and can change them, and the tasks and comments your teammates add to it are sent to Jira from your Atlassian account. Once in Kabiri, synced tasks are part of your workspace like any other task, so the Kabiri Assistant can see them when you ask it about your work, and, when you ask it to, change them for you (add, edit, move or complete a task, or comment on it), only as you could yourself; those changes reach Jira like any other change you make in Kabiri. Your Jira data is never used to train, fine-tune, test or evaluate any model, ours or anyone else's. Stopping. When you stop syncing a project, Kabiri takes it out of its site's webhook (and removes the webhook when no synced project is left on the site), and its tasks stay in Kabiri as ordinary tasks unless you choose to remove them; nothing is deleted in Jira. When you disconnect Jira from Connected apps, we stop syncing every project, remove the webhooks, and delete the stored tokens and sync records straight away. Atlassian gives apps no way to withdraw their own access, so to take Kabiri off your Atlassian account as well, remove it in your Atlassian account settings, under Connected apps (id.atlassian.com/manage-profile/apps). When you close your Kabiri account, the tokens and the rest are deleted with your account data. Who is responsible. Kabiri is the controller of the copy of your Jira data held in Kabiri, independently of Atlassian. This policy covers how Kabiri handles it; Atlassian's privacy policy does not. If we ever have a security breach that touches data from Jira, we will tell Atlassian as well as you. Inside Jira, Atlassian's terms and privacy policy apply. Atlassian (Atlassian Pty Ltd in Australia, and Atlassian US, Inc. in the United States; privacy@atlassian.com) hosts Jira Cloud data in the United States unless your organisation has chosen another region, and Kabiri's requests go through Atlassian's API in the United States, so data you sync is processed outside the UK / EU. For transfers from the UK, Atlassian relies on the EU-US Data Privacy Framework and its UK Extension, and on Standard Contractual Clauses with the UK Addendum where those don't apply. We never use your Jira data for advertising, sell it, or share it with anyone else. Jira is a trademark of Atlassian. Kabiri isn't affiliated with or endorsed by Atlassian.
• Linear (optional two-way task sync): Kabiri works with Linear through its official API, to keep the Linear projects you choose in step with Kabiri projects, both ways (a sync set up before 4 November 2026 may keep a whole Linear team instead: what is said here of a project's issues then covers the team's). Each Linear project becomes a Kabiri project, and its team's workflow states become the project's columns. When you make a new Kabiri project and ask for it, Kabiri also creates a Linear project with the same name, in the team you pick, and syncs it. Nothing syncs until you connect your Linear account and agree, on the connect screen, to Kabiri reading the issues in the projects you pick, with their comments and the names of the people on them (and their email addresses, where their Linear profiles show them), keeping a copy up to date in Kabiri, and making the changes you and your team make in Kabiri in Linear too: creating, editing, moving, commenting on and deleting those issues. Kabiri writes to Linear only the changes made in Kabiri to the projects you sync, and only as you. One Linear workspace can be connected to a Kabiri account at a time. When you connect, Linear asks you to let Kabiri read the workspace you choose, change its issues, and create issues, projects and comments; Kabiri uses that access only for the projects you sync and, when you ask, for the Kabiri Assistant's search (below). A Linear connection made before 2 November 2026, when Linear synced one way only, stays read-only: Kabiri keeps reading from Linear but sends it nothing until you reconnect Linear and agree to the two-way sync on the connect screen. Kabiri stores the connection itself: the access and refresh tokens Linear issues, encrypted and never sent back to your browser, your Linear user id and name, so Connected apps can show which account is connected, and the name of the workspace. What our servers receive. To let you choose what to sync, our servers ask Linear for the teams and projects in your workspace (their names, keys and ids). For each project you sync, they ask Linear for its team's workflow states, labels and issues: each issue's title, description, state, due date, priority, estimate, labels, cycle, project and parent, who it is assigned to and who subscribes to it (their names and Linear user ids, and their email addresses where their Linear profiles show them), the issues it is related to, its comments (with their authors' names and ids) and the names and links of its attachments. Linear tells our servers when an issue or a comment changes, and our servers check each synced project every few hours in case a notice went missing. From those notices we look only at which issue or comment changed, then ask Linear for the latest version. We keep nothing from the projects you don't sync, apart from the alerts below. Alerts from Linear. Unless you switch them off in Settings > Notifications, our servers check your Linear notifications (the ones Linear shows you in its inbox) about every 5 minutes for anything that mentions you, is assigned to you, or is a new comment on something of yours, including outside the projects you sync. For each one, Kabiri keeps a short alert for 30 days: the name of the person who did it, what kind it was, the issue's key, title and place, and its link into Linear. Kabiri never keeps a comment's text for an alert, and never writes anything to Linear for one. Searching Linear with the Kabiri Assistant. When you ask the Kabiri Assistant to find something in Linear, our servers search the issues of every team your Linear connection can see, not only the ones you sync, and read back the matching issues' keys, titles, states, the names of the people they are assigned to, priorities, teams and when they last changed; for one issue you ask about, also an extract of its description and its three newest comments with their authors' names. Never an email address, and nothing is written to Linear. What comes back is sent to the Google Gemini API, our processor, for that one reply, as with any question you ask the Kabiri Assistant. Our servers don't store the search results: they are used for that reply, which you see in the chat like any other answer, and the only thing stored is a count of how many Linear searches you made in the hour. What we keep. For each Linear project you sync, Kabiri keeps its issues as tasks in the Kabiri project you chose: their titles, descriptions (with their formatting; anything Kabiri can't show is kept on our servers so it isn't lost when you edit the description), states (the project's columns are its team's own workflow states), due dates, priorities, labels, estimates and other values (as Kabiri fields), who each is assigned to and who subscribes to it (by name), the issues each is related to, whether each is done, and their links into Linear, along with the project's name, states and fields. People in the team. Our servers keep the names and Linear user ids of the people named on the synced issues (assignees, subscribers, the authors of comments), and their email addresses when their Linear profiles show them. The tasks hold their names; the email addresses are kept on our servers only, never shown in the app or sent anywhere. They are used for one thing: to tell whether that person is you or, on a Kabiri team project, one of your teammates, by the email address their Kabiri account has verified. Nobody else is looked up. That record on our servers is deleted when you stop syncing the project, disconnect Linear or close your account. When a project stops syncing or you disconnect Linear in Kabiri, the tasks Kabiri keeps no longer hold Linear people's names or ids: the assignees and subscribers from Linear are taken off them, and comments by Linear people, and mentions of them, show as "Linear user". Linear gives apps no way to report that someone has closed their account, so a person's name stays on the project's synced tasks until you stop syncing or disconnect. These tasks are stored with the rest of your workspace in the UK / EU (europe-west2), with a small record of what was last synced, so that a change on either side reaches the other. Changes you make to those tasks in Kabiri (adding, editing, completing or deleting them, moving them between the team's workflow states, their field values, who they are assigned to, who subscribes to them and what they are related to) are sent to Linear, as your Linear permissions allow; an issue deleted from Kabiri goes to the trash in Linear. When Kabiri creates an issue, it may add a link to kabiri.co to it, which only Kabiri reads, so a retry never makes the issue twice. Each task links to its issue in Linear. Repeats are not synced. Comments. The comments on synced issues are kept in step both ways too. Comments written in Linear, including by people who don't use Kabiri, are kept with the task in Kabiri: their text (with its formatting), when each was written or last edited, and the display name of whoever wrote it. Comments written on a synced task in Kabiri are posted to Linear from your Linear account; one a teammate writes starts "From" and their name, so Linear shows who said it. You can edit or delete your own comments in Kabiri and Linear follows, and edits and deletions made in Linear show in Kabiri; other people's comments can't be changed in Kabiri. Files attached to an issue in Linear are shown by name, with a link that opens them in Linear; nothing is copied into Kabiri's storage, and files you attach in Kabiri are not sent to Linear. Only people who can see the Kabiri project see its comments and attachments. They are deleted with their task, and when you stop syncing a project or disconnect Linear and choose to remove its tasks; tasks you keep keep their comments as ordinary Kabiri comments, shown as written by "Linear user", and their attachments as links into Linear. If you sync a team in a Kabiri team workspace, everyone who can see that project sees its Linear tasks and can change them, and the tasks and comments your teammates add to it are sent to Linear from your Linear account. Once in Kabiri, synced tasks are part of your workspace like any other task, so the Kabiri Assistant can see them when you ask it about your work, and, when you ask it to, change them for you (add, edit, move or complete a task, or comment on it), only as you could yourself; those changes reach Linear like any other change you make in Kabiri. Your Linear data is never used to train, fine-tune, test or evaluate any model, ours or anyone else's. Stopping. When you stop syncing a project, its tasks stay in Kabiri as ordinary tasks unless you choose to remove them; nothing is deleted in Linear. When you disconnect Linear from Connected apps, we stop syncing every project, revoke Kabiri's access at Linear, and delete the stored tokens and sync records straight away. You can also remove Kabiri yourself in Linear, under Settings > Account > Authorized applications: Linear then tells our servers, which stop syncing and delete the stored tokens and the list of people on the synced issues. The tasks already in Kabiri stay as they are, with the names on them, until you stop syncing the project or disconnect Linear in Connected apps. A workspace admin can also remove Kabiri. When you close your Kabiri account, the tokens and the rest are deleted with your account data. Who is responsible. Kabiri is the controller of the copy of your Linear data held in Kabiri, independently of Linear Orbit, Inc., the company that makes Linear. This policy covers how Kabiri handles it; the Linear privacy policy does not. Inside Linear, its terms and privacy policy apply. Linear Orbit, Inc. is a US company, and Kabiri's requests go through its API, so data you sync may be processed outside the UK / EU. We never use your Linear data for advertising, sell it, or share it with anyone else. Linear is a trademark of Linear Orbit, Inc. Kabiri isn't affiliated with or endorsed by Linear.
• ClickUp (optional two-way task sync): Kabiri works with ClickUp through its official API, to keep the ClickUp lists you choose in step with Kabiri projects, both ways. Each list becomes a Kabiri project, and the list's statuses become the project's columns. When you make a new Kabiri project and ask for it, Kabiri also creates a ClickUp list with the same name, in the space and folder you pick, and syncs it. Nothing syncs until you connect your ClickUp account and agree, on the connect screen, to Kabiri reading the tasks in the lists you pick, with their comments and the names of the people on them (and their email addresses, where ClickUp shows them to apps), keeping a copy up to date in Kabiri, and making the changes you and your team make in Kabiri in ClickUp too: creating, editing, moving, commenting on and deleting those tasks. Kabiri writes to ClickUp only the changes made in Kabiri to the lists you sync, and only as you. When you connect, ClickUp asks you which of your workspaces Kabiri may reach; Kabiri uses that access only for the lists you sync and, when you ask, for the Kabiri Assistant's search (below). A ClickUp connection made before 2 November 2026, when ClickUp synced one way only, stays read-only: Kabiri keeps reading from ClickUp but sends it nothing until you reconnect ClickUp and agree to the two-way sync on the connect screen. Kabiri stores the connection itself: the access token ClickUp issues, encrypted and never sent back to your browser, your ClickUp user id and name, so Connected apps can show which account is connected, and the names of the workspaces you let Kabiri reach. What our servers receive. To let you choose what to sync, our servers ask ClickUp for the spaces, folders and lists in those workspaces (their names and ids). For each list you sync, they ask ClickUp for its statuses, its custom fields and its tasks: each task's name, description, status, start and due dates, priority, tags, time estimate, custom field values, its parent task, who it is assigned to and who watches it (their names and ClickUp user ids, and their email addresses where ClickUp shows them to apps), the tasks it depends on or is linked to, its comments (with their authors' names and ids) and the names and links of its attachments. For each list you sync, Kabiri sets up one webhook at ClickUp, for that list only, so that ClickUp tells our servers when a task or a comment in it changes, and our servers check each synced list every few hours in case a notice went missing. From those notices we look only at which task or comment changed, then ask ClickUp for the latest version. We keep nothing from the lists you don't sync, apart from the alerts below. Alerts from ClickUp. Unless you switch them off in Settings > Notifications, our servers check the tasks assigned to you and their new comments about every 5 minutes for anything that mentions you, is assigned to you, or is a new comment on something of yours, including outside the projects you sync. For each one, Kabiri keeps a short alert for 30 days: the name of the person who did it, what kind it was, the task's key, title and place, and its link into ClickUp. A comment's text is read only to see whether it mentions you. Kabiri never keeps a comment's text for an alert, and never writes anything to ClickUp for one. Searching ClickUp with the Kabiri Assistant. When you ask the Kabiri Assistant to find something in ClickUp, our servers look through the most recently changed tasks of up to two of your ClickUp workspaces, not only the lists you sync, and read back the matching tasks' titles, statuses, the names of the people they are assigned to, priorities, lists and when they last changed; for one task you ask about, also an extract of its description and its three newest comments with their authors' names. Never an email address, and nothing is written to ClickUp. What comes back is sent to the Google Gemini API, our processor, for that one reply, as with any question you ask the Kabiri Assistant. Our servers don't store the search results: they are used for that reply, which you see in the chat like any other answer, and the only thing stored is a count of how many ClickUp searches you made in the hour. What we keep. For each ClickUp list you sync, Kabiri keeps its tasks as tasks in the Kabiri project you chose: their names, descriptions (with their formatting; anything Kabiri can't show is kept on our servers so it isn't lost when you edit the description), statuses (the project's columns are the list's own statuses), start and due dates, priorities, tags, time estimates and other values (as Kabiri fields), who each is assigned to and who watches it (by name), the tasks each depends on or is linked to, whether each is done, and their links into ClickUp, along with the project's name, statuses and fields. When a ClickUp task has several assignees, the first is the task's assignee in Kabiri and the others are shown among its watchers, marked as assigned in ClickUp. People in the list. Our servers keep the names and ClickUp user ids of the people named on the synced tasks (assignees, watchers, the authors of comments), and their email addresses when ClickUp shows them to apps. The tasks hold their names; the email addresses are kept on our servers only, never shown in the app or sent anywhere. They are used for one thing: to tell whether that person is you or, on a Kabiri team project, one of your teammates, by the email address their Kabiri account has verified. Nobody else is looked up. That record on our servers is deleted when you stop syncing the project, disconnect ClickUp or close your account. These tasks are stored with the rest of your workspace in the UK / EU (europe-west2), with a small record of what was last synced, so that a change on either side reaches the other. Changes you make to those tasks in Kabiri (adding, editing, completing or deleting them, moving them between the list's statuses, their field values, who they are assigned to, who watches them and what they depend on or link to) are sent to ClickUp, as your ClickUp permissions allow. Each task links to its task in ClickUp. Repeats are not synced. Comments. The comments on synced tasks are kept in step both ways too. Comments written in ClickUp, including by people who don't use Kabiri, are kept with the task in Kabiri: their text, when each was written or last edited, and the display name of whoever wrote it. Comments written on a synced task in Kabiri are posted to ClickUp from your ClickUp account; one a teammate writes starts "From" and their name, so ClickUp shows who said it. You can edit or delete your own comments in Kabiri and ClickUp follows, and edits and deletions made in ClickUp show in Kabiri; other people's comments can't be changed in Kabiri. Files attached to a task in ClickUp are shown by name, with a link that opens them in ClickUp; nothing is copied into Kabiri's storage, and files you attach in Kabiri are not sent to ClickUp. Only people who can see the Kabiri project see its comments and attachments. If you sync a list in a Kabiri team workspace, everyone who can see that project sees its ClickUp tasks and can change them, and the tasks and comments your teammates add to it are sent to ClickUp from your ClickUp account. Once in Kabiri, synced tasks are part of your workspace like any other task, so the Kabiri Assistant can see them when you ask it about your work, and, when you ask it to, change them for you (add, edit, move or complete a task, or comment on it), only as you could yourself; those changes reach ClickUp like any other change you make in Kabiri. Your ClickUp data is never used to train, fine-tune, test or evaluate any model, ours or anyone else's. Stopping. ClickUp tasks don't stay behind in Kabiri. When you stop syncing a list, Kabiri removes its webhook and deletes the list's tasks from Kabiri, with their comments, the fields that came from ClickUp and the record of the people on them; they stay in ClickUp, and nothing is deleted there. When you disconnect ClickUp from Connected apps, we do the same for every synced list, and delete the stored token and sync records straight away. ClickUp gives apps no way to withdraw their own access, so to take Kabiri off your ClickUp account as well, remove it in ClickUp, under Settings > Apps. You can also start there: ClickUp doesn't tell apps when they are removed, so the next time our servers reach ClickUp and are turned away, the sync pauses and Connected apps asks you to reconnect ClickUp. If you don't reconnect within 7 days, Kabiri deletes the synced tasks, their comments, the stored token and the list of people, as it does when you disconnect. When you close your Kabiri account, the synced ClickUp tasks are deleted, in team projects too, and the token and the rest are deleted with your account data. Who is responsible. Kabiri is the controller of the copy of your ClickUp data held in Kabiri, independently of Mango Technologies, Inc., the company that makes ClickUp. Kabiri, not ClickUp, is responsible for the privacy, security and integrity of the ClickUp data Kabiri holds. This policy covers how Kabiri handles it; the ClickUp privacy policy does not. If we ever have a security breach that touches data from ClickUp, we will tell ClickUp as well as you. Inside ClickUp, its terms and privacy policy apply. Mango Technologies, Inc. is a US company, and Kabiri's requests go through its API, so data you sync may be processed outside the UK / EU. We never use your ClickUp data for advertising, sell it, or share it with anyone else. ClickUp is a trademark of Mango Technologies, Inc. Kabiri isn't affiliated with or endorsed by ClickUp.
• HubSpot (optional two-way CRM sync): Kabiri works with HubSpot through its official API, to keep the contacts, companies and deals in the HubSpot account you connect in step with your Kabiri Contacts, both ways. Nothing syncs until you connect your HubSpot account and agree, on the connect screen, to Kabiri reading those records, keeping a copy up to date in Kabiri, and making the changes you make in Kabiri in HubSpot too: creating, editing and deleting contacts, companies and deals and moving deals between stages. Kabiri writes to HubSpot only the changes made in Kabiri to synced records, and only as you. One HubSpot account can be connected to a Kabiri account at a time. When you connect, HubSpot asks you to let Kabiri read and write your contacts, companies and deals and read their property definitions; Kabiri uses that access only for the sync and, when you ask, for the Kabiri Assistant's search (below). Kabiri stores the connection itself: the access and refresh tokens HubSpot issues, encrypted and never sent back to your browser, your HubSpot account id and its domain, so Connected apps can show which account is connected, and the email address of the HubSpot user who connected. What our servers receive. For each contact: name, email addresses, phone numbers, job title, address, website, and the company HubSpot links it to. For each company: name, domain, website, industry, phone, address and description. For each deal: name, amount and currency, pipeline and stage, close date, description, and the contact and company it is with; and the account's deal pipelines and their stages. HubSpot tells our servers when a record changes (once that is switched on), and our servers check the account every 15 minutes (every 6 hours on Free) in case a notice went missing. From those notices we look only at which record changed, then ask HubSpot for the latest version. Searching HubSpot with the Kabiri Assistant. When you ask the Kabiri Assistant to find something in HubSpot, our servers search the account's contacts, companies and deals and read back the matching records' names, a line of context (a job title, a domain, a deal's state) and their links into HubSpot. What comes back is sent to the Google Gemini API, our processor, for that one reply, as with any question you ask the Kabiri Assistant; it is not stored beyond a count of how many HubSpot searches you made in the hour. What we keep. The synced records, as contacts, companies and deals in your Kabiri Contacts, each carrying its HubSpot id and link; the pipelines and stages; and the sync's own record of which Kabiri record mirrors which HubSpot one and what was last agreed between the two. Stopping and deleting. Stop syncing keeps the records in Kabiri as ordinary ones and removes their HubSpot ids and links, or deletes them if you ask; disconnecting does the same and revokes Kabiri's access at HubSpot, deleting the tokens. Closing your Kabiri account deletes all of it. Nothing from HubSpot is ever used to train or evaluate a model, used for advertising, sold, or shared with anyone else. HubSpot, Inc. is a US company, and Kabiri's requests go through its API, so data you sync may be processed outside the UK / EU. HubSpot is a trademark of HubSpot, Inc. Kabiri isn't affiliated with or endorsed by HubSpot.
• Zoho CRM (optional two-way CRM sync): Kabiri works with Zoho CRM through its official API, to keep the contacts, accounts and deals in the Zoho CRM organisation you connect in step with your Kabiri Contacts, both ways. Nothing syncs until you connect your Zoho account and agree, on the connect screen, to Kabiri reading those records, keeping a copy up to date in Kabiri, and making the changes you make in Kabiri in Zoho too: creating, editing and deleting contacts, accounts and deals and moving deals between stages. Kabiri writes to Zoho only the changes made in Kabiri to synced records, and only as you. One Zoho CRM organisation can be connected to a Kabiri account at a time. When you connect, Zoho asks you to let Kabiri read and change your contacts, accounts and deals, read your pipelines, stages and field settings, read your organisation's name and your own user record, and set up notifications; Kabiri uses that access only for the sync and, when you ask, for the Kabiri Assistant's search (below). Kabiri stores the connection itself: the access and refresh tokens Zoho issues, encrypted and never sent back to your browser, your organisation's id and name, your email address, and which of Zoho's data centres your organisation is in (the US, Europe, India, Australia, Japan, Canada, Saudi Arabia or China), so that every request goes to that centre. What our servers receive. They ask Zoho for your contacts (names, email addresses, phone numbers, job titles, postal addresses, the account each belongs to, a Twitter handle, notes), your accounts (names, websites, industries, phone numbers, addresses, notes) and your deals (names, amounts and currencies, stages and pipelines, the contact and account each is with, closing dates, notes), and your pipelines and stages. Zoho tells our servers when a record changes (a notification naming the record, never its contents), and our servers check for changes every 15 minutes on paid plans and every 6 hours on Free in case a notice went missing. Searching Zoho CRM with the Kabiri Assistant. When you ask the Kabiri Assistant to find something in Zoho, our servers search your organisation's contacts, accounts and deals and read back the matching records' names, companies, job titles, stages and links. Never an email address, and nothing is written to Zoho. What comes back is sent to the Google Gemini API, our processor, for that one reply, and the only thing stored is a count of how many searches you made in the hour. What we keep. Your synced records, as contacts, companies and deals in your Kabiri Contacts, with their links into Zoho; the matching between each Kabiri record and its Zoho record; and your Zoho pipelines and stages. Stopping the sync keeps the records in Kabiri without their Zoho links, or removes them if you ask; disconnecting does the same and revokes Kabiri's access at Zoho, deleting the tokens. Closing your Kabiri account deletes all of it. Nothing from Zoho is ever used to train or evaluate a model, used for advertising, sold, or shared with anyone else. Zoho Corporation is a US company with data centres around the world, and Kabiri's requests go through its API in the centre your organisation is in, so data you sync may be processed outside the UK / EU. Zoho and Zoho CRM are trademarks of Zoho Corporation. Kabiri isn't affiliated with or endorsed by Zoho.
• Pipedrive (optional two-way CRM sync): Kabiri works with Pipedrive through its official API, to keep the people, organisations and deals in your Pipedrive company in step with your Kabiri Contacts, both ways. Nothing syncs until you connect your Pipedrive account and agree, on the connect screen, to Kabiri reading your Pipedrive persons, organisations, deals and pipelines, keeping a copy up to date in Kabiri, and making the changes you and your team make in Kabiri in Pipedrive too: creating, editing, moving between stages and deleting those records. Pipedrive stays the record: where the two sides disagree, Pipedrive's version is kept. Kabiri writes to Pipedrive only the changes made in Kabiri to synced records, and only as you. When you connect, Pipedrive asks you to let Kabiri read and change your contacts and deals and manage its own webhooks; Kabiri uses that access only for the sync and, when you ask, for the Kabiri Assistant's search of your Pipedrive. Kabiri stores the connection itself: the access and refresh tokens Pipedrive issues, encrypted and never sent back to your browser, your Pipedrive company id, its name and its address on Pipedrive (the host Kabiri's requests go to), and your email address as Pipedrive gives it, so Connected apps can show which company is connected. What our servers receive. Our servers ask Pipedrive for your pipelines and stages (their names and ids) and for your persons, organisations and deals: each person's name, email addresses, phone numbers, job title, address and organisation; each organisation's name and address; each deal's title, value and currency, stage, the person and organisation it is with, its expected close date and whether it is open, won or lost; with each record's Pipedrive id and the time it last changed. They keep these as contacts, companies and deals in your Kabiri contact book (or your team's, if you chose it), with the Pipedrive ids beside them, and a record of what was last agreed between the two sides, so a change on one side can be told from a change on both. Kabiri reads nothing else from Pipedrive: not your notes, activities, emails, files, products or leads. Pipedrive tells our servers when a person, organisation or deal changes through three webhooks Kabiri registers in your account, protected by a password of their own; those messages name the record and carry its new values, which are used only to update the copy. On Free, the sync fills your contacts up to Free's limits and keeps the rest waiting, on our servers, until there is room. The Kabiri Assistant can, when you ask, search your Pipedrive for people, organisations and deals that are not in Kabiri yet; it reads names and ids only and keeps nothing. Stopping, disconnecting and deleting. Stop syncing keeps the synced records in Kabiri as ordinary contacts, companies and deals, with their Pipedrive ids and links removed, unless you ask for them to be deleted from Kabiri; Kabiri's webhooks and its record of the sync are deleted either way. Disconnect does the same for everything, removes the webhooks, revokes Kabiri's access at Pipedrive and deletes the stored tokens and the connection. To take Kabiri off your Pipedrive account as well, remove it in Pipedrive, under Settings > Tools and apps > Marketplace apps. When you close your Kabiri account, the tokens and the sync's records are deleted with your account data, and the contacts, companies and deals with it. Who is responsible. Kabiri is the controller of the copy of your Pipedrive data held in Kabiri, independently of Pipedrive OÜ, the company that makes Pipedrive. Kabiri, not Pipedrive, is responsible for the privacy, security and integrity of the Pipedrive data Kabiri holds. This policy covers how Kabiri handles it; the Pipedrive privacy policy does not. If we ever have a security breach that touches data from Pipedrive, we will tell Pipedrive as well as you. Inside Pipedrive, its terms and privacy policy apply. Pipedrive OÜ is an Estonian company, inside the EU; your company's data may be held in the EU or the US, where Pipedrive keeps it, and Kabiri's requests go to your company's own Pipedrive address. We never use your Pipedrive data for advertising, sell it, share it with anyone else, or use it to train, fine-tune, test or evaluate any model. Pipedrive is a trademark of Pipedrive OÜ. Kabiri isn't affiliated with or endorsed by Pipedrive.
• Salesforce (optional two-way CRM sync, coming later): Once it ships, Salesforce will work the same way as HubSpot, Pipedrive and Zoho CRM above: nothing syncs until you connect your Salesforce org and agree, on the connect screen, to Kabiri reading its contacts, accounts and opportunities, keeping a copy up to date in your Kabiri Contacts, and making the changes you make in Kabiri in Salesforce too, as you. Your sign-in is kept sealed on Kabiri's servers, never in the browser; stopping the sync or disconnecting leaves the records in your Contacts without Salesforce's ids (or deletes them, if you ask) and deletes the stored tokens; closing your Kabiri account revokes Kabiri's access at Salesforce. Salesforce, Inc. is a US company, so data you sync may be processed outside the UK / EU. We never use your Salesforce data for advertising, sell it, share it with anyone else, or use it to train, fine-tune, test or evaluate any model. Salesforce is a trademark of Salesforce, Inc. Kabiri isn't affiliated with or endorsed by Salesforce.
• Weather and Currency (dashboard widgets): These two widgets fetch public data straight from your browser, so each service below sees your IP address and browser details, as any website you visit would. The Weather widget sends your location to Open-Meteo for the forecast and to OpenStreetMap's Nominatim service to name your town. It only does this after you allow location access or choose a city; it never guesses your location from your IP address. The coordinates it sends and saves are rounded to about one kilometre, and anything you type into the city search goes to Open-Meteo to find matching places. Your chosen city and its rounded coordinates are saved to your Kabiri account so the widget shows the same place on your other devices. The Currency widget asks ExchangeRate-API and Frankfurter for exchange rates; it sends them the currencies you picked and nothing about you. These services may process requests outside the UK / EU. If you would rather none of this happened, remove the widgets from your dashboard and no request is made.
• Other Kabiri users you share your calendar with: Where your plan includes calendar sharing, you decide who can see your calendar. A member of your Kabiri team who adds it sees it straight away, unless you turn off "Teammates see my calendar without asking" in Calendar settings, when they ask you first. Anyone else always asks, and you approve or decline each request. Someone you share with sees the titles, times, locations, descriptions and guest lists of the events on calendars you own. Events you mark private, events on iCloud calendars, and calendar feeds (ICS links) you subscribe to show only as "Busy". Your events are read from your connected calendars when they look, and are kept on our servers for no more than about two minutes. Their browser keeps a copy so your calendar opens quickly, and drops it when the share ends or they sign out. We keep a record of each share (who asked, whether it was approved, and when) so that it can be enforced. You can see everyone who has access, and remove them, under Calendar settings › Sharing, and anyone you share with can stop at any time. A team share ends if either of you leaves the team, and every share ends when either of you closes your account. When someone asks for a calendar by email address, we never tell them whether that address has a Kabiri account.
• Legal Requirements: We may disclose your information if required to do so by law or in the good faith belief that such action is necessary to comply with a legal obligation, protect our rights or property, or protect the safety of users or the public.

4. Security of Your Information

We use administrative, technical, and physical security measures designed to help protect your personal information. Your data is stored securely using Google Cloud Platform's Firebase suite. While we have taken reasonable steps to secure the personal information you provide, no security measures are perfect or impenetrable.

5. Data Retention

We retain your personal data only for as long as necessary to fulfil the purposes for which it was collected. When you delete your account, all associated personal data is permanently deleted within 30 days, or straight away if you choose immediate erasure, except as set out below. Until then you can reopen the account by signing in again. Waitlist data is retained until you are invited to the Service or until you request removal. A project you delete is kept in Deleted projects for 30 days, so that you can restore it, and is then permanently deleted with its tasks; you can delete it for good sooner from there. Daily product usage counts (recorded only if you accept analytics cookies) are deleted after 90 days. Scheduled briefings are deleted automatically 7 days after they are written, and the Briefing's memory of what each day's briefing covered (subjects, sender and people's names, and the titles of meetings and tasks, never message text) keeps only the last 7 days briefed. A device's notification token is kept until you sign out on it, the browser stops accepting it, or you close your account. The record of the files people send you through a file request is kept while the request exists, and deleted with the request or your account (see File requests). Aggregated, anonymised analytics data may be retained indefinitely for statistical purposes.

Two exceptions, for work you share with a team. Files and messages you have shared into a shared team or workspace conversation are part of that shared history that other members rely on, so they are not automatically removed when you delete your account. Comments you have written on tasks in a team's projects stay with that team in the same way, but without your name: when your account is permanently deleted, your name is removed from them and they are shown as written by a "Former member". Comments on tasks in your own projects are deleted with your account. Comments other people wrote in Todoist or Asana, copied onto a team's tasks, stay with those tasks, with their authors' names. Comments copied from Jira stay too, but without their authors' names: when a project stops syncing or Jira is disconnected (as it is when you close your account), the tasks Kabiri keeps no longer hold Jira people's names or ids, and those comments show as written by "Jira user" (see Jira). Comments copied from Linear are kept the same way, and show as written by "Linear user" (see Linear). Tasks and comments copied from ClickUp don't stay behind: they are deleted from Kabiri, team projects included, when the list stops syncing, ClickUp is disconnected or you close your account (see ClickUp). If you need shared content removed as well, contact our Data Protection Officer at privacy@kabiri.co and we will handle it in line with your rights and the legitimate interests of the other team members.

Cached data stored locally in your browser (see Section 1, "Local Browser Storage") is retained on your device until you clear your browser's site data for kabiri.co, sign out, or your browser evicts it under its own storage limits. Account deletion removes data on our servers; it does not reach into other devices to clear local caches there.

6. Your Rights

You may request access to, correction of, or deletion of your personal data at any time by contacting our Data Protection Officer at privacy@kabiri.co. You can also export a copy of your data and permanently delete your account from your account settings within the app. For full details of your rights under GDPR, see our GDPR Compliance page.

If you believe we have not handled your personal data properly, you have the right to lodge a complaint with the Information Commissioner's Office (ICO), the UK supervisory authority, at ico.org.uk. We would, however, appreciate the chance to address your concerns before you approach the ICO.

7. Children's Privacy

Kabiri is intended for professional and business use and is not directed at children. You must be at least 18 years old to use the Service. We do not knowingly collect personal data from anyone under 18. If you believe a child has provided us with personal data, please contact our Data Protection Officer at privacy@kabiri.co and we will delete it.

8. Changes to This Policy

We may update this Privacy Policy from time to time. We will notify you of any changes by posting the new Privacy Policy on this page and updating the "Last Updated" date. You are advised to review this Privacy Policy periodically.

9. Contact Us

If you have questions about this Privacy Policy, please contact our Data Protection Officer at privacy@kabiri.co, or reach us at hello@kabiri.co.