Privacy Policy
Effective: October 1, 2026
This Privacy Policy explains how Noto Scribe (“we”, “us”, “our”) collects, uses, stores, and shares information when you use our service across the web, iOS, Android, macOS, and Windows: recording and notes, shared spaces, group chat and direct messages, huddles (voice and video calls), pages, and Meher, our AI teammate. By using Noto Scribe you agree to the practices described below.
If you only read one section, read “How we use AI services” below. It explains exactly what leaves your device and which third-party providers process it.
1. Who we are
Noto Scribe transcribes conversations in real time and turns them into structured, searchable notes. It also gives teams a place to work on them together: shared spaces with group chat, direct messages, voice and video huddles, and pages. It is operated by Meta Multiples Private Limited, a company incorporated in India, which is the data controller for the information described in this policy.
Our mobile apps’ identifier is com.metamultiples.notoscribe, and our desktop app’s is com.notoscribe.app. For privacy questions, contact us at support@metamultiples.in, or write to us at 201, Rizvi Chambers-1, Hill Road, Bandra West, Mumbai, Maharashtra, India – 400050.
2. Information we collect
Account information
- Email address and authentication credentials
- If you sign in with Google or Apple, the name and email address they share with us
- Display name and the preferences you set, such as language, time zone, and notification and email settings
- Authentication tokens managed by our auth provider (Supabase)
Content you create or add
- Audio recordings you capture, and recordings or videos you upload
- Transcripts derived from those recordings, including word-level timestamps and speaker labels
- AI-generated notes, titles, summaries, and action items
- Photos you add to a recording, and screen recordings you make in the web or desktop app
- Files, photos, videos, documents, and links you add to a space, import, or share into Noto Scribe from another app
- Pages and databases, with their comments, suggested edits, version history, and uploaded files
- Messages you send in a space’s group chat or in direct messages, with their attachments, reactions, GIFs, and link previews
- Your conversations with Meher, in writing or by voice, including the audio and transcript of voice sessions
- Meher’s memory of you: a short profile Meher keeps from your voice sessions so she can pick up where you left off. You can view, edit, or clear it in Settings.
- Spaces and subspaces, and the metadata you attach to them
Huddles (voice and video calls)
- Your voice, and your camera and screen when you turn them on, go through our call server to the other people in the huddle.
- Huddles are recorded. Our call server records the voices of everyone in a huddle, including guests and Meher, but not cameras. When a member shares their screen from the web or desktop app, their browser also records the share and uploads it. The call shows that it is recorded. Afterwards, anyone in the space can turn the recording into a note with a transcript and summary.
- Details of each huddle: who started it, who joined and when, its topic, and who turned notes on
- People can join a huddle as a guest through a link, without an account. We ask a guest only for a name, which the others in the call see and which appears in the huddle’s transcript.
Activity in shared spaces
- The members of each space, with their names, email addresses, and roles, and the email addresses of people invited who haven’t joined yet
- Which messages you’ve read, your notification settings for each chat, and the messages you’ve saved
- Whether you’re online or typing, and where your cursor is when you edit a page with others. These are shown live and not stored.
- The pages you’ve viewed recently, and daily page-view counts, which you can turn off with “Count my page views” in Settings
- The notifications we’ve sent you, and whether you opened them
Payment information
When you buy recording credits, you pay through Razorpay’s checkout. You enter your card, UPI, or other payment details there; we never receive or store your full card number. We keep a record of each payment (order and payment IDs, the pack you bought, amount and currency, payment method type, country, status, and invoice number), along with your credit balance and recording usage.
Device and usage information
- IP address, user agent, operating system, and app version
- Approximate location (country, region, and city), which our hosting provider works out from your IP address and we use to show prices in your currency and to suggest languages; and your time zone
- On phones, the app’s push notification token (and, on iPhone, a separate token for huddle calls), with the platform and app version
- Crash logs and diagnostic events from our mobile apps
- Update checks from our mobile apps, which include the device model, operating system, language and region settings, time zone, and an identifier for the app on that device
- Timestamps of feature usage
- On macOS and Windows, the desktop application can capture system audio (for example, to transcribe a video call). System audio is captured only while you have explicitly started a recording and is never accessed in the background.
- To offer to record when a meeting starts, the desktop application checks, on your computer, whether a meeting app such as Zoom, Microsoft Teams, Webex, FaceTime, or Slack is running and, on a Mac, whether a Google Meet tab is open in Chrome. This check runs while the app is open and stays on your computer: we don’t receive or store the list of your apps or tabs.
Storage on your device
On the web, your browser keeps your sign-in session and preferences, such as your theme and unsent chat drafts, in local storage. While you record, it also keeps the audio in its local database (IndexedDB) so that a recording survives a crash or a lost connection; those copies are deleted once the recording is saved. We use a small number of functional cookies, such as one that remembers whether the sidebar is open, and no advertising or analytics cookies. You can clear all of this by clearing your browser’s site data for Noto Scribe.
Our mobile apps keep your sign-in tokens in the device’s secure storage, and keep recordings, uploads that haven’t gone through yet, and a list of your spaces on the device so they work offline. Files you share into the app from another app are copied into the app’s own storage so they can be added to a space.
3. How we use your information
- To provide the recording, transcription, and note-taking service
- To deliver what you share: show your messages, notes, and pages to the people you share them with, connect huddles, and ring the people you call
- To generate AI-assisted titles, summaries, action items, translations, search results, and Meher’s answers — this involves sending your content to third-party AI service providers, as described in “How we use AI services” below
- To send push notifications and emails: sign-in codes, password resets, receipts, invitations, and the digests and alerts you choose in Settings (see “Notifications and emails” below)
- To sell recording credits, keep billing records, and track your usage
- To investigate abuse, debug issues, and enforce rate limits that protect service availability
- To comply with applicable law and respond to lawful requests from public authorities
We do not sell your personal information, and we do not use your recordings, transcripts, notes, messages, calls, or pages to train AI models. Our agreements with AI sub-processors also prohibit them from using your content to train their models.
4. How we use AI services
Noto Scribe relies on a small set of third-party AI service providers to power its core features. Before your recordings or notes are processed for the first time in our mobile apps, you are shown an in-app disclosure (“Your data & AI”) summarizing what this section describes, and you must affirmatively tap “I agree” before any data is sent. You can re-read that disclosure any time from Settings → Privacy & AI.
What we send, and to whom
| Data sent | Provider | Purpose |
|---|---|---|
| The audio of your recordings (in short pieces while you record, and in full afterwards), recordings and videos you upload, podcast episodes you import, huddle recordings, and what you say to voice search | ElevenLabs (speech-to-text) | Producing a transcript with word-level timestamps and speaker labels |
| Transcripts and note content, pages, database rows, and files and photos you give Meher | Google (Gemini API) | Generating titles, summaries, action items, translations, re-transcription cleanup, in-note AI edits, Meher’s answers, space search results, database fields, and Meher’s memory of you |
| Photos you add to recordings, still frames from screen recordings and huddle screen shares, images and scanned PDFs you import, and videos you import from YouTube, Instagram, or TikTok | Google (Gemini API) | Describing what they show, so it can be included in notes and search |
| Text, images, and PDFs from your notes, pages, chat messages, and files, and what you type into search | Google (Gemini API) | Building the search index you search your spaces with |
| When Meher answers in a chat: the conversation, the space’s summary, members’ names and roles, and matching passages from the space’s notes and messages (private notes are left out) | Google (Gemini API) | Writing Meher’s reply |
| A search query Meher writes from the conversation, when she looks something up on the web for you | Google (Gemini API with Google Search) | Answering with current information from the web |
| Your microphone audio, while Live Translate is open during a recording, sent straight from your device | Google (Gemini API) | Showing a live translation on screen; it isn’t saved |
| Your voice, while you talk with Meher by voice, with your name, the space’s name, recent note titles and summaries, your previous session’s transcript, Meher’s memory of you, and your local time | ElevenLabs (voice agent), which uses Google’s Gemini API to write Meher’s replies | Letting Meher listen and answer aloud. ElevenLabs keeps the session’s audio for up to 30 days; we keep the recording and transcript as a note. |
| Voices in a huddle, only while someone in it has turned on spoken translation, and only the voices being translated | Speech translation provider (Google’s Gemini API) | Translating what is said, by voice, into another language for the people in the call |
| Voices in a huddle, only while someone in it has brought in Meher, the AI teammate, with the space’s summary and its members’ names and roles | Speech-to-text and voice provider (ElevenLabs); large language model provider (Google’s Gemini API); message classification provider (TypeSafe, through OpenRouter) | Letting Meher follow the call, decide when she is being spoken to, and answer aloud |
| Up to 30 recent messages of a space’s group chat, with senders’ names and file names, each time someone posts a message that doesn’t mention Meher (never direct messages) | Message classification provider (TypeSafe, through OpenRouter, or Google’s Gemini API) | Deciding whether Meher should reply to or react to a message she wasn’t tagged in |
Huddle translation is off unless someone turns it on. When they do, the voices being translated are sent to Google’s Gemini API to translate them; anyone whose voice is being translated is told so in the call. We don’t store the translated audio, and it is never part of the huddle’s recording, transcript or notes.
Meher joins a huddle only when someone in it brings her in. She leaves when she is removed or asked to leave, when the call ends, or when the recording time of the person who brought her in runs out. What she says in a huddle is part of its recording. In a space’s group chat she reads new messages to decide whether to reply or react, as a teammate would; she doesn’t do this in direct messages.
Rules our AI sub-processors operate under
Each AI sub-processor receives only the data necessary to return a result and is bound by a written data-processing agreement requiring them to:
- Process your data only on our documented instructions
- Apply confidentiality, encryption in transit, and equivalent technical and organizational safeguards
- Retain your data only as long as needed to return a result, and delete it after processing
- Not use your content to train or improve their AI models
- Notify us of any security incident affecting your data without undue delay
We assess each provider for protections equivalent to those described in this Privacy Policy before adding them to the service. The providers named in this policy are the ones we use today, and we update this policy when that changes. Questions about them are welcome at support@metamultiples.in.
Connecting Noto Scribe to other AI assistants
You can connect an outside AI assistant, such as Claude or ChatGPT, to your Noto Scribe account. When you approve the connection, the assistant can, acting as you, list your spaces, read and search your notes, transcripts, summaries, images, and pages (search also covers chat messages, files, and links in your spaces), and query databases. If you also allow page editing, it can create and edit pages, add and update database rows, and press buttons on pages. It can’t reach anything you can’t. What the assistant does with that content is governed by its provider’s terms and privacy policy, not ours. You can disconnect an app at any time from Settings.
Sharing notes with outside AI assistants (clipboard)
Noto Scribe offers a Use with LLMaction that bundles a note’s title, summary, and full transcript into a single markdown prompt and copies it to your device clipboard, so you can paste it into a third-party AI assistant (for example, ChatGPT or Claude). This action is user-initiated, and nothing is sent anywhere until you paste it. Once content leaves your clipboard for another application, that application’s privacy terms govern what happens to it— not ours.
5. Sharing with other people
Much of Noto Scribe is for working with others, so some of what you add is visible to other people by design:
- Spaces.Recordings and notes you add to a shared space (unless you mark them private), and the space’s files, links, pages, and group chat, are visible to its owner and members, including people who join later. Owners and admins of a space can also open its subspaces, including private ones, but never direct messages or anyone’s private notes and pages.
- Members.Members of a space can see each other’s names, email addresses, and roles.
- Direct messages. Messages and huddles in a direct message are visible to everyone in that conversation. People added to a group conversation later can see its earlier messages.
- Huddles.The people in a huddle, including guests, see and hear you live. For 30 days afterwards, anyone in the space can play the huddle’s recording and make a note from it, even if they weren’t in the call.
- Meher.Meher’s replies in a shared chat or huddle are visible to everyone there, and can draw on what the space holds, except private notes and pages.
- Public links.You can create a link to a note or a page that anyone with the link can open without signing in. A note link shows the note’s write-ups and, unless you turn them off, its transcript and audio; it shows its screen recording only if you turn that on. Public links work until they are turned off, and we ask search engines not to index them.
- Invitations.When you invite someone by email, we store their address and send them an invitation that shows your name and what they’re invited to.
- Leaving a space. If you leave a space or are removed from it, the messages, notes, pages, and files you shared there stay with the space for its members. You keep your own recordings and notes, and your private conversation with Meher in that space is deleted.
6. Notifications and emails
Our mobile apps send push notifications for new messages, mentions, replies, and huddles. They are delivered through Google Firebase Cloud Messaging (and, for huddle calls on iPhone, Apple’s push notification service), and can include the sender’s name, the space or conversation, and a preview of the message. You can turn previews off in the app’s Settings. Only one phone per account gets notifications and huddle calls: the one you signed in on most recently, or the one where you tapped “Use this phone”. We keep a history of the notifications we send you. On the web, your browser shows notifications only while Noto Scribe is open.
We send emails through Resend: sign-in codes and password resets, invitations to spaces and pages, low-balance and top-up messages, and, unless you turn them off, a receipt after each note, a daily digest, a daily email about unread chat messages (with short message previews), and emails about mentions, replies, and suggested edits on pages and weekly database digests. You can turn the optional ones off in Settings → Notifications.
7. Sub-processors and third parties
Beyond the AI providers described above, we rely on the following sub-processors to operate the service. Each receives only the data necessary to perform its function and is bound by a written data-processing agreement.
| Vendor | Purpose | Data shared |
|---|---|---|
| Supabase | Authentication, database, live updates (chat, presence, and editing pages together), and storage for some older recordings | Account data and the content you create |
| Cloudflare (R2) | File storage | Recordings, huddle recordings, photos, screen recordings, and the files you upload or share |
| Vercel | Web hosting and APIs | Standard request metadata, including IP address and approximate location, and server logs |
| Oracle Cloud | Hosting our call server for huddles, in India | Huddle audio, video, and screen shares; copies of huddle recordings for up to 7 days |
| Resend | Transactional email delivery | Email address, message content |
| Razorpay | Payments | Payment details you enter in its checkout, the order, your email address, and your account ID |
| Google Firebase Cloud Messaging | Push notifications on Android and iPhone | Push token and notification content |
| Apple Push Notification service | Huddle calls on iPhone | Call token, the caller’s name, the space, and your account ID |
| Sentry | Crash and error reports from the mobile apps | Device and app details, error details, and the IDs of the note or space involved |
| Stallion | Delivering updates to the mobile apps | Device model, operating system, language and region, time zone, app version, and an identifier for the app on that device |
| Apify | Importing YouTube, Instagram, and TikTok links | The link you import |
| Jina AI and Firecrawl | Reading web pages you import | The link you import |
Other services that receive limited data
These services aren’t our sub-processors, but they see limited data when you use the features that rely on them:
- GIPHYprovides the GIFs in chats and huddles. When you search for one, our server sends GIPHY your search words, not who you are. GIFs load from GIPHY’s servers, which receive the viewer’s IP address.
- Embeds on pages(for example YouTube, Vimeo, Loom, Figma, Google Docs, Drive and Maps, GitHub Gist, CodePen, Miro, Typeform, X, and Spotify) load in the viewer’s browser from those services, which receive the viewer’s IP address and may set their own cookies. HTML blocks on pages can load code and fonts from public sources such as cdnjs, jsDelivr, unpkg, and Google Fonts, and our emoji load from Google Fonts.
- Link previews. When a link is posted in a chat or bookmarked on a page, our server visits it to fetch its title and preview image, so the site sees our server rather than you.
- Imports.When you import a podcast episode or a Google Doc, our servers fetch it from where it’s published, such as Apple Podcasts, Spotify, the podcast’s feed, or Google Docs.
- Google and Apple, when you choose to sign in with them, under their own privacy policies.
- GitHub, which hosts our desktop app’s updates; the app checks there for new versions.
- AI assistants you connect, as described in “Connecting Noto Scribe to other AI assistants” above.
8. Recording responsibility
Laws governing the recording of conversations vary by jurisdiction. In some places only one party to a conversation must consent to a recording; in others, all parties must consent. You are responsible for understanding and complying with the laws that apply to you, including obtaining any necessary consent before recording. Noto Scribe is a tool; it does not determine the lawfulness of a recording you choose to make.
This applies to huddles too. Every huddle is recorded, and the call shows that it is. Before you invite people into a huddle, or bring in Meher or spoken translation, make sure that everyone in the call knows and that recording it is lawful where you and they are.
9. Data retention
- Recordings, transcripts, notes, pages, messages, and files are kept until you delete them (or, in a shared space, someone with permission does), the space they belong to is deleted, or your account is deleted.
- Huddle recordings, voices and screen shares alike, are deleted 30 days after the call; our call server keeps its copy for up to 7 days. A note made from a huddle keeps its audio, transcript, and a record of who spoke when, like any other note.
- When a message is deleted, its text and attachments are removed and a placeholder shows where it was. Page trash is emptied after 30 days, and page history after 90 days unless a version is bookmarked.
- Live translations, in recordings and in huddles, are never stored.
- Push tokens are deleted when you sign out, when another phone takes over your notifications, or when they stop working. Connected apps keep access until you disconnect them.
- When you delete your account in the app, it is deactivated at once and you are signed out. You can restore it by signing back in within 90 days; after that, we delete your personal data from live systems. When you ask us by email to delete your account, we remove your personal data from live systems within 30 days of verifying the request.
- Messages, pages, and comments you posted in spaces that other people use stay there for them, no longer linked to your account. Spaces you own are deleted with your account, along with everything in them, unless you hand them over to another member first.
- Backups containing deleted data are purged on a rolling 90-day cycle.
- Limited operational logs (security, abuse, billing where applicable) may be retained for up to 12 months.
10. Your rights
Depending on where you live, you may have the right to access, correct, export, restrict the processing of, or delete your personal data, and to object to certain processing. Residents of the European Economic Area, the United Kingdom, and California have additional rights under the GDPR, UK DPA, and CCPA respectively. These rights also apply if we hold your data without you being a Noto Scribe user, for example as a huddle guest or someone invited by email.
You can also manage much of this yourself: delete recordings, notes, messages, and pages; mark notes private; turn public links off; choose which emails and notifications you get and whether they show message previews; turn off “Count my page views”; view, edit, or clear Meher’s memory of you; and disconnect connected AI assistants.
To exercise any of these rights, email support@metamultiples.in or follow the steps on our data deletion page. We will respond within 30 days.
11. Children
Noto Scribe is not directed to children under 13 (or under 16 in the European Economic Area). We do not knowingly collect personal information from children. If you believe a child has provided us with personal information, contact us and we will delete the account.
12. International transfers
We are based in India, and our call server for huddles runs in India. Your data may also be processed in other countries where our providers operate, including the United States. Where required by law, we put appropriate safeguards in place (such as Standard Contractual Clauses) with our sub-processors to protect transferred data.
13. Security
We protect your data with industry-standard measures, including TLS in transit, encryption at rest at our storage providers, scoped API credentials, and access controls. Messages and huddles are encrypted in transit but are not end-to-end encrypted: our servers process them to deliver, record, transcribe, and search them, and to let Meher take part. No method of transmission or storage is perfectly secure, but we work continuously to protect your information.
14. Changes to this policy
We may update this Privacy Policy from time to time. When we make material changes we will notify you by email or through an in-app notice and update the “Effective” date at the top of this page.
This version, effective October 1, 2026:
- Covers group chat, direct messages, huddles, pages, push notifications, and connecting other AI assistants
- Names the providers we use and what each one receives
- Explains account deletion, including the 90-day window in which you can restore your account
15. Contact
Privacy questions, requests, and complaints: support@metamultiples.in.