Every tool in this category now does the same three things: joins or listens to your meeting, produces a summary with action items, and lets you search past conversations. That convergence happened a while ago, which means picking one on features alone no longer tells you much.
The differences that remain are about how the tool captures the meeting, who it is really built for, and what it does with the transcript afterward.
The bot-in-the-call versus bot-free split
This is the most consequential technical difference and it is easy to miss because most product pages bury it.
- Bot-in-the-call — a visible participant joins your Zoom, Meet or Teams call to record. Fireflies's core capture method works this way, and it is the most reliable option for cross-platform consistency, but it announces itself to everyone on the call, including external guests who may not expect it
- Bot-free — the tool listens through your computer's audio instead of joining as a participant. Granola is built around this from the ground up, and Fathom, Otter and Notta all now offer a bot-free desktop option alongside their bot-based one
- Bot-free matters most for external meetings, sales calls with cautious counterparts, or any context where a visible recording participant changes how people talk. It matters less on internal team calls where everyone already expects to be recorded
Granola (8.4) — for people who take their own notes anyway
Granola's whole design starts from not joining the call at all, and it layers AI on top of the notes you were already going to take rather than replacing your attention with a transcript. It supports Zoom, Meet, Teams and in-person conversations via mobile, and its Granola Chat lets you query meeting history in plain language — "what did I promise to do this week" is the example the product leads with.
The detail worth flagging for this audience specifically: Granola connects to Claude, Cursor and Replit via MCP, which makes it the one tool on this list built with an eye toward the same agent ecosystem the rest of this site covers, rather than treating meetings as a closed silo.
Fathom (8.3) — for sales teams that live in follow-up
Fathom's pitch leans hard on speed to action — its own claim is three times faster from meeting insight to next step, and six-plus hours saved weekly on follow-up work. AI Scorecards add performance coaching specifically for sales conversations, which is a feature the general-purpose tools on this list do not really compete on.
If your primary use case is sales calls that need to become CRM entries and follow-up emails with minimal manual work, this is the one built closest to that exact workflow.
Otter (8.2) — the straightforward, well-priced default
Otter's pricing is the most transparent on this list — Business at $19.99 a month per user for 6,000 monthly transcription minutes is a number you can budget against immediately, where several competitors make you talk to sales to find out. Its AI Chat queries both meetings and connected apps, and channel organisation groups meetings by team or project.
This is the reasonable default when you do not have a specialised need pulling you toward one of the others — accurate, clearly priced, unglamorous.
Fireflies (8.1) — for teams that need compliance guarantees
Fireflies leads with security credentials the others mention less prominently: SOC 2 Type II, GDPR and HIPAA compliance, plus an explicit zero-data-retention-for-AI-training guarantee. AskFred answers questions across your meeting history, and Live Assist gives real-time coaching suggestions during a call rather than only afterward.
If your organisation has a compliance team that needs a specific, citable answer about data handling before approving a tool, Fireflies makes that answer easiest to find.
tl;dv (7.9) — for teams that need cross-meeting pattern-spotting
tl;dv's multi-meeting analysis aggregates insights across many conversations rather than treating each one in isolation, which is aimed at sales coaching use cases — tracking playbook adherence and flagging coaching opportunities across a rep's call history rather than any single call.
It is worth picking specifically when the question you are asking is about a pattern across meetings, not the content of one meeting.
Notta (7.7) — for the cheapest real usage tier and 58 languages
Notta's paid tier starts at $8.17 a month billed annually, which undercuts most of this list. Its 58-language transcription support and real-time bilingual translation are the standout feature for teams working across languages rather than English-only teams choosing on note quality alone.
Read AI (7.4) — for whoever wants an assistant, not just a notetaker
Read AI positions itself furthest from "just transcribe my meetings" — Ada, its digital assistant, extends into email summarisation and task automation, and it is available as a Claude Connector, a ChatGPT app and an MCP server, which puts it in the same agent-adjacent category as Granola rather than the pure-notetaker group.
The lowest score here reflects breadth without matching depth in any one area yet — worth watching rather than the first pick today.
How to actually choose
Skip the feature checklist and ask three questions instead.
- Do external people join your calls? If yes, weigh bot-free capture seriously — Granola, or the bot-free modes in Fathom, Otter and Notta
- Is this primarily for sales, or general team meetings? Sales-heavy use pulls toward Fathom or tl;dv; general use is served well by Otter or Fireflies
- Does your organisation need a specific compliance answer before approving anything? Start with Fireflies's credentials page rather than a feature comparison