Otter.ai

Otter.ai is one of the most recognizable names in AI transcription, and the product has only grown more ambitious. The company has refined its real-time transcription engine for years, and live meetings still surface as one of its strengths: Otter transcribes conversations as they happen, identifies speakers, and produces searchable, shareable transcripts teams can work on together. In early 2026 Otter crossed $100M ARR and by July 2026 it was claiming more than 35 million users, and it has since repositioned itself as a “Conversational Knowledge Engine” rather than a notetaker, framing meetings as raw material for a cross-team, cross-time knowledge layer.

The platform integrates with Zoom, Google Meet, and Microsoft Teams, automatically joining meetings via a bot that records and transcribes. Otter also offers desktop apps for Mac and Windows that can capture audio directly without a bot joining the call. The bot remains the default and is standard across the industry, but the desktop app gives users a bot-free alternative. For teams that have normalized AI recording, the bot is a non-issue. For client-facing calls or sensitive discussions, the desktop app’s bot-free mode reduces friction.

The bigger story is what Otter has built on top of transcription. Its headline product is now the AI Meeting Agent suite — Meeting Agent, Sales Agent, and SDR Agent — a set of voice-activated, real-time participants rather than passive notetakers. Speak to the agent during a call and it can pull answers from your company-wide meeting database, schedule follow-ups, draft emails, coach sales reps with live objection handling, and even run product demos autonomously. The older OtterPilot notetaking workflow still exists underneath all this, but Otter’s center of gravity has clearly shifted toward agentic, conversational AI.

Key Features

Otter’s core strength is its real-time transcription engine, which delivers live captions with speaker identification during meetings. Independent 2026 testing puts everyday accuracy around 85% — strong and consistent on standard US English, though weaker on non-native accents, jargon, and technical terms. The live transcript itself remains a highlight.

On top of that sits the AI Meeting Agent suite. The agents are voice-activated and answer questions mid-meeting from across your entire meeting history, not just the current call. The Sales Agent and SDR Agent extend this into revenue workflows with live coaching, follow-up scheduling, and email drafting. In July 2026 Otter added Live Assist, an Enterprise-only real-time coaching agent that shows tip cards only the user can see. After each meeting, Otter still generates summaries and action items, and the chat feature lets you query your meetings in natural language (now metered, from roughly 20 queries a month on free up to 200 on Business).

Otter’s MCP (Model Context Protocol) server, now available even on the free Basic plan, lets you pipe meeting data into external AI models, including Claude. The public API and webhooks, which connect Otter to proprietary CRMs and internal systems, remain Enterprise-only. Alongside the built-in Salesforce and HubSpot integrations and the shared collaborative workspace, this gives Otter a genuinely deep knowledge-connection and integration story.

Pricing

Otter’s free Basic plan includes 300 transcription minutes per month, but with real limits: each conversation is capped at 30 minutes, you get only 3 lifetime file imports, your history is limited to the 25 most recent conversations, and AI Chat is restricted to roughly 20 queries per month.

The Pro plan is $16.99 per month, or $8.33 per month billed annually ($99.99 per year, about 51% off). It raises the allowance to 1,200 minutes a month, lifts the per-conversation cap to 90 minutes, allows 10 imports per month, and adds advanced search, custom vocabulary, priority support, and 50 AI Chat queries a month. The Business plan runs $30 per user per month, or $19.99 per user billed annually (about 33% off), and adds unlimited meeting transcription, a 4-hour per-conversation cap, 200 AI Chat queries a month, plus admin controls, analytics, and compliance features. Enterprise pricing is custom. Since August 2026 Otter also offers a 50% student discount.

Per-user pricing adds up for larger teams. A team of 10 on the Business plan pays $300 per month at the monthly rate. That is significant when privacy-focused alternatives like Hedy offer comparable capability at a lower price point.

Privacy & Security

This is where Otter requires careful consideration. All audio is streamed to Otter’s cloud servers for processing, so meeting content passes through third-party infrastructure. Otter encrypts data in transit and at rest and holds SOC 2 Type II compliance — but two issues go beyond generic cloud-architecture caution.

First, by default Otter trains its AI models on de-identified user data. Its privacy page confirms the practice but does not describe a self-serve opt-out; Enterprise accounts are reportedly opted out by default. Second, and more consequentially, the federal class action against Otter cleared its first real hurdle in August 2026.

The case began as Brewer v. Otter.ai and was consolidated in October 2025 into In re Otter.AI Privacy Litigation (5:25-cv-06911, N.D. Cal.). It alleges that Otter’s Notetaker and OtterPilot recorded private conversations without all-party consent and used those recordings to train its models. On 13 August 2026, Judge Eumi K. Lee granted Otter’s motion to dismiss in part and denied it in part. The claims under the federal Electronic Communications Privacy Act, California’s Invasion of Privacy Act, the Illinois Biometric Information Privacy Act, unjust enrichment, and California’s Unfair Competition Law all survived. The CFAA, CDAFA, and Washington Privacy Act claims were dismissed, as were the common-law privacy claims of every plaintiff but one.

The reasoning is what makes this ruling matter beyond Otter. In holding that Otter could be a third-party eavesdropper under CIPA section 631, the court leaned on the fact that Otter independently collects, retains, and uses meeting communications for its own commercial purposes rather than merely passing them through on a customer’s behalf. That distinction — tool versus independent participant — does not turn on anything unique to Otter’s product. Any assistant that retains recordings and uses them for its own benefit, model training very much included, is exposed to the same argument.

None of this is a finding of liability. Surviving a motion to dismiss means the plaintiffs pleaded enough to proceed to discovery, nothing more. But it converts what was a filed complaint into a live case with a published holding, and it is a documented reason for organizations in regulated industries, or anyone handling confidential client information, to scrutinize how Otter’s defaults align with their data-governance requirements. On 17 September 2026 Otter filed its answer, denying that it creates voiceprints and denying that it failed to obtain consent; the plaintiffs’ class-certification motion is due 13 May 2027. Our notetaker lawsuit tracker follows the case.

Best For

Otter.ai is best suited for teams where collaboration around meeting content is a priority and cloud processing fits the security posture. Marketing teams, product teams, and sales organizations that want live, agentic AI in the room will get the most from it — the Sales Agent and SDR Agent in particular are built for revenue workflows. It is a weaker fit for legal teams, executive discussions, or any context where meeting content is highly confidential and the default training on user data and ongoing litigation are dealbreakers.

How Otter.ai Compares to Hedy

Otter has years of market presence, a polished collaboration layer, and now a genuinely capable agentic suite — voice-activated agents, live sales coaching, an MCP server, and a public API on Enterprise. Hedy does not try to match that breadth feature for feature. Where Hedy pulls decisively ahead is architecture: with Local AI enabled on supported hardware, its AI runs end-to-end on-device. Transcription, summaries, and real-time coaching all happen locally, so your meeting content never leaves your device, while Otter streams everything to the cloud and trains on de-identified user data by default. Otter’s bot-based recording works with supported video platforms (its desktop app adds a bot-free option), but Hedy captures system audio directly with no bot at all and works across any meeting tool, phone calls, and in-person conversations. Otter has no offline capability; Hedy works fully offline with Local AI. For teams where privacy is non-negotiable, Hedy offers what Otter structurally cannot. For teams that want Otter’s collaborative annotation and in-meeting agents and accept cloud processing, Otter remains a strong contender.