Skip to main content
otter.ai

Is Otter.ai Safe? Storage, Training, and Meeting Bots

Vemoir Team
Vemoir Team Mac meeting records by Aurum Avis Labs · Oberägeri, Switzerland
12 min read

Is Otter.ai safe for your meetings depends on what you mean by safe: who sees a bot in the call, where audio and transcripts are stored, whether content is used to improve models, and how deletion works. Otter publishes detailed privacy, security, and help pages. This article summarizes what those documents say as of October 2026. It is practical guidance, not legal advice, and it does not say whether Otter fits your contract, sector rules, or internal policy.

If you need a shared cloud notetaker with calendar sync and real-time team transcripts, Otter can be a reasonable choice when participants consent and your organization accepts U.S. cloud processing. If the audio must stay on your Mac unless you export it, a local recorder is the different model. Both can be valid. The question is which risks you want to reduce.

Facts at a glance (from Otter’s own pages)

TopicWhat Otter documents
Where data is storedCloud compute and storage with Amazon Web Services in the United States, including SSE on S3 buckets (Privacy Policy, Privacy & Security)
Who runs the serviceOtter.ai, Inc. (Delaware, U.S.). Enterprise customers may be the data controller under their own agreement (Privacy Policy)
Bot visibilityOtter Notetaker joins as a visible participant; Otter states it cannot join anonymously (Notetaker overview)
Model trainingPrivacy Policy lists training proprietary AI on de-identified audio and transcriptions; Privacy & Security says de-identification before training and no manual human review for that training path (Privacy Policy, Privacy & Security)
Third-party AI backendsOtter states customer data is not used to train AI service providers’ models sent through their APIs (Privacy & Security)
DeletionDelete moves conversations to Trash for 30 days, then permanent server deletion; empty Trash deletes immediately (Help: delete a conversation, Privacy & Security)
EncryptionAES-256 server-side encryption on AWS storage (Privacy & Security)
Human accessSupport access to audio or transcript requires explicit customer consent before manual review (Privacy & Security)

Policies change. Check the links above before you rely on them for a decision.

Is Otter.ai safe? What that usually means for a meeting notetaker

Most professionals asking is Otter.ai safe are really asking four things:

  1. Visibility: Will clients or colleagues see an AI participant?
  2. Location: Does my recording leave my device and sit in a vendor cloud?
  3. Reuse: Can the vendor use my voice or text to train models or share with subprocessors?
  4. Control: Can I delete copies and limit who sees the transcript afterward?

Otter’s answers boil down to this: it is a visible cloud service with documented training on de-identified content, plus sharing controls and deletion flows. That can work for many internal standups. It can be a poor fit for confidential client work, air-gapped policies, or teams that forbid guest bots.

Where Otter stores recordings and transcripts

Otter is a cloud service. The Privacy Policy (effective June 16, 2026) describes Meeting and Uploaded Information, including audio recordings, transcripts, and related usage data, processed to provide the service. Under With whom we share, Otter names cloud service providers including Amazon Web Services, based in the United States.

The Privacy & Security page adds that Otter uses AWS S3 with server-side encryption (SSE) and AES-256. That describes technical safeguards on Otter’s side. It does not by itself answer whether your employer or client allows U.S. hosting for meeting content. That remains your organization’s decision.

If your employer buys an enterprise plan, the Privacy Policy says that enterprise customer is the data controller and its own privacy policy applies to the processing of your personal information.

Does Otter train on your recordings?

This is the detail many privacy reviews focus on.

Privacy Policy: One stated purpose is to improve and monitor the Services, including training our proprietary AI technology on de-identified audio recordings and on transcriptions (which may contain Personal Information). Otter also describes data labeling service providers who create training and evaluation data for product features.

Privacy & Security FAQ: Otter says it uses a proprietary method to de-identify user data before training, that training is automatic (audio and transcripts not manually reviewed by a human for that training path), and that training data is encrypted.

Terms of Service (section 18.4): Otter may use aggregated and/or deidentified data derived from use of the services for purposes including machine learning and training, in a form that does not identify you.

In plain words, Otter documents model improvement using de-identified meeting content. It does not promise that your raw recordings never influence its product AI. Otter separately states that AI service providers used through APIs do not use customer data to train their models (Privacy & Security). That is a different layer from Otter’s own proprietary training.

If you need no vendor-side training on meeting audio at all, read Otter’s current settings and enterprise terms, or choose a workflow where audio never enters Otter’s cloud.

Who sees the bot in your meeting?

Otter’s meeting assistant is Otter Notetaker (formerly OtterPilot). According to Otter’s Notetaker overview:

  • Notetaker joins Zoom, Google Meet, and Microsoft Teams as a participant.
  • The display name appears as your profile first name plus “Notetaker (Otter.ai)”.
  • Otter states Notetaker cannot join anonymously or without appearing to other participants.
  • Guest participants must be allowed in the meeting, or Notetaker may be blocked.

The Privacy & Security page says recording starts when you or a meeting participant presses Record or invites Otter to the meeting. Calendar sync and auto-join are optional product features described in Otter help articles.

For client calls, visibility matters as much as storage. A visible bot is easier to disclose than a hidden recorder, but many IT policies still block guest notetakers entirely.

Calendar auto-join and workspace defaults

Otter can join from Google or Microsoft calendar sync when you enable it. Before a sensitive call, open Otter settings and check auto-join, share with workspace, and email summary defaults. Otter help articles describe turning auto-join off per meeting or globally. Enterprise admins can set organization-wide rules, per Otter pricing.

If you stay on Otter, treat those toggles as part of your security review, not an afterthought.

Deletion, retention, and shared copies

User deletion: Otter’s help article Delete a conversation states that deleting a conversation unshares it immediately, moves it to Trash for 30 days, then permanently deletes it from Otter’s servers. Emptying Trash deletes sooner. Only the conversation owner can delete a shared conversation.

Policy retention: The Privacy Policy says Otter stores personal information as long as necessary for the purposes in the policy or as required by law, and describes making deleted personal information irrecoverable when deleted.

De-identified data: The Privacy Policy treats deidentified or aggregated data separately from personal information. Deleting a conversation does not automatically tell you whether derived de-identified datasets are removed from every downstream use. If you need certainty for a compliance review, ask Otter or your enterprise admin rather than assuming.

Honest limits (what Otter does not solve for you)

On Privacy & Security, Otter cites a SOC 2 Type 2 report, GDPR and CCPA standards, HIPAA requirements, and the ISO 27001/2 framework as the basis of its security policies (a framework, not an ISO certificate). Those pages describe Otter’s program. They do not guarantee that your use case is compliant, privileged, or approved by a client. Nothing in this article certifies Otter or your recording choices.

Other limits to keep in mind:

  • Cloud jurisdiction: U.S. AWS storage may conflict with data-handling rules your client imposes, independent of Otter’s marketing copy.
  • Subprocessors: The Privacy Policy lists categories such as analytics, advertising partners, data labeling, and AI service providers. Review Otter’s subprocessor list for your plan.
  • Third-party voices: Recordings can contain other people’s speech. Consent and retention remain your responsibility (Otter’s Privacy & Security page also reminds users to follow local recording laws).
  • Sharing: Transcripts can be shared inside Otter. A deletion on your account does not remove what a colleague exported to email or Slack.
  • Litigation: As of October 2026, Otter is a defendant in a consolidated U.S. class action, In re Otter.AI Privacy Litigation (N.D. California). Plaintiffs allege that Notetaker recorded meetings without all participants’ consent and that recordings were used for AI training. On August 13, 2026, the court let several claims go forward and dismissed others. The allegations are unproven.

When Otter is often the better choice

Otter is strong when you want a shared, cloud-native notetaker:

  • Automatic join from Google or Microsoft calendar with team-visible transcripts.
  • Real-time transcription and Otter’s in-product summaries, chat, and slide capture features (see Otter’s product pages).
  • Collaboration inside Otter workspaces without building your own pipeline.
  • Organizations that already approved Otter’s vendor risk profile and U.S. processing.

If your team lives in Otter today and legal signed off, asking is Otter.ai safe may mean tightening settings (auto-join rules, sharing defaults, trash hygiene) rather than switching tools.

When a local Mac workflow fits better

Choose local capture when keeping audio off vendor servers by default matters more than calendar bot convenience:

  • Vemoir for Mac records microphone and system audio on your Mac, with no bot in the participant list (how it works, supported meeting apps).
  • Transcription and summaries run on the Mac with local models unless you opt into a cloud model with your own API key (capture and models).
  • For an audio file you already have after the call, the free transcribe recording tool runs in the browser tab without uploading the file to Vemoir.

Otter optimizes for shared cloud notes. Vemoir optimizes for local processing and explicit export.

If your next call cannot include a guest notetaker, you can record the meeting on your Mac with Vemoir instead of inviting Otter.

Compare cloud bot vs local capture

Otter Notetaker (Otter docs)Vemoir on Mac (product facts)
Joins call as participantYes, visible guestNo bot; you start and stop recording
Default processingOtter cloud (AWS U.S.)On-device models; cloud models opt-in with your key
OfflineRequires service connectivity for cloud featuresFull workflow can run offline after models install
Best whenTeam wants shared cloud transcriptsAudio should stay on the Mac unless you export

Try Otter with open eyes, or keep audio on your Mac

Staying on Otter: Review Privacy & Security, your auto-join settings, and Trash retention before client calls. Tell participants a notetaker is present, and align with your internal policy.

Avoiding a cloud notetaker: Use the Vemoir Mac app to capture the next Teams, Zoom, or Meet call locally, or transcribe an existing file with transcribe recording without uploading audio to Vemoir.

Frequently asked questions

Is Otter.ai safe for confidential client calls?

Otter stores content in U.S. cloud infrastructure and documents AI training on de-identified recordings and transcripts. Whether that is acceptable for a given client contract is your organization’s call. Otter also shows a visible Notetaker participant, which helps disclosure but does not replace consent or policy review.

Does Otter sell my meeting data for ads?

Otter’s Privacy & Security page says Otter does not sell personal information to third parties and serves no third-party ads in its products. The Privacy Policy separately lists advertising partners and notes that some disclosures may count as a “sale” or targeted advertising under certain U.S. state laws. Read both sections for your jurisdiction.

Can Otter join without anyone noticing?

No. Otter’s Notetaker overview says Notetaker cannot join anonymously and appears in the participant list.

Where are Otter recordings stored?

Otter names Amazon Web Services in the United States for cloud storage and compute in its Privacy Policy and describes AWS S3 with SSE on Privacy & Security.

Does Otter use my recordings to train AI?

Otter’s Privacy Policy includes training proprietary AI on de-identified audio recordings and transcriptions. Privacy & Security describes de-identification before training and no manual human review on that automatic training path.

How do I delete Otter conversations?

Use Otter’s in-app delete. Help docs say deleted items sit in Trash for 30 days, then are permanently removed from Otter’s servers, or delete immediately from Trash (help article).

Is Otter encrypted?

Otter documents AES-256 server-side encryption on AWS storage (Privacy & Security). The same page says two-factor authentication is available on all plans.

What is a local alternative without a meeting bot?

Vemoir for Mac records system audio and microphone locally during the call, with no bot in the participant list. For a file you already have, use the transcribe recording tool in the browser without uploading audio to Vemoir.

Does this article prove GDPR or nLPD compliance for Otter?

No. This article summarizes Otter’s public documentation and general product tradeoffs. It is not legal advice and does not evaluate Swiss, EU, or other statutory compliance for your situation.

Next step

Read Otter’s current privacy policy and security page before your next recorded call. If you decide cloud bots are out of scope, open the Vemoir Mac app or the free transcription tools to keep processing on your Mac.

otter.ai privacy meeting notes
Vemoir Team

Written by

Vemoir Team

Mac meeting records by Aurum Avis Labs · Oberägeri, Switzerland

We build Vemoir, a Mac app for local-first meeting records: capture from the menu bar, transcribe on your Mac, and turn calls into minutes and action items.

Customize your cookie preferences. Essential cookies cannot be disabled as they are required for the website to function properly.

Essential Cookies

Required for basic website functionality, security, and the free in-browser tools weekly limit (week id and count only).

Always active

Analytics Cookies

Help us understand how visitors interact with our website to improve user experience. Includes Google Analytics and Microsoft Clarity session recordings.

Marketing Cookies

Used to track visitors across websites to display relevant and engaging advertisements.