Every figure here carries the date it was read and the source it came from. How scoring works →
AI Note Takers for Meetings, Scored on What a Seat Costs and What the Contract Lets You Keep
2 tools measured · Vouch Score data collected 11 October 2026 · highest composite first
An AI note taker records a meeting, turns the speech into a transcript and writes a summary of it, usually by sending a bot into the call on Zoom, Google Meet or Microsoft Teams, sometimes by capturing audio on your own machine instead. The buyer is rarely one person for long. It starts with somebody who wants their own sales calls or interviews written up, and it spreads to a team the moment those notes are shared, which is where most of the cost appears.
What decides the fit is less the transcript than everything around it: what a seat costs once billing is monthly rather than annual, how much recording the plan keeps before it stops, whether the notes can leave the product in a format you can edit, who the contract says owns the recording of a call you hosted, and who the contract makes responsible for asking the other people on the call.
The cards here measure those things from the vendor's own documents and from review platforms whose sampling can be read off the page. Where no independent measurement of transcription accuracy exists, the capability position carries a named criterion under its own name rather than a vendor's accuracy claim.
Top score: Fireflies.ai. It holds the highest Vouch Score composite on this page, 83.1/100, measured 23 September 2026. The label is computed from the cards, not chosen by us, and it moves the day another tool measures higher. Read what each card measured before you take it as advice for your own work.
1FFireflies.aitop score · highest composite83.1/100 · high
Individuals and small teams who want every call transcribed and summarised without changing how they meet
Sends a bot named Fred into Zoom, Google Meet and Teams calls to record, transcribe and summarise them, billed per seat, with a documented API and exports you can edit.
Fireflies.ai sends a bot, Fred, into Zoom, Google Meet and Teams calls to record, transcribe and summarise them. It scores 83.1; the capability row counts what a paid seat lets you take away, not accuracy. Pro costs $18 per seat a month billed monthly and $10 a month billed annually, the view the pricing page loads in. Fireflies' help article, read 22 September 2026, says Fred is set by default to join every meeting with a web-conference link.
Visit Fireflies.ai ↗Read the full review →Free tier: 'Free forever', with 'Unlimited transcription*', '400 mins of storage/team' and '20 AI credits'. Paid tiers read $18 (Pro) and $29 (Business) per seat a month in the monthly view and $10 and $19 per seat a month in the annual view, which is the state the page loads in. Enterprise is $39 per seat a month and sold annually only. Both views read on 2026-09-22.
Individuals who want unlimited recording and transcription on a free plan, and teams that want call notes pushed into a CRM
Records, transcribes and summarises Zoom, Google Meet and Teams calls, with or without a bot in the room, and gives every plan the API.
Fathom is an AI meeting notetaker from Fathom Video Inc. that records Zoom, Google Meet and Microsoft Teams calls and returns a transcript and a summary. Its pricing page, read 11 October 2026, prints a 90 day guarantee on every plan; its Terms, last updated March 4, 2026, say all fees are non-refundable. Premium is $20 a month billed monthly and $16 a month billed annually. The 78.9 composite holds no accuracy measurement.
Visit Fathom ↗Read the full review →Free plan: 'Free forever', with 'Unlimited recordings + transcriptions'; advanced summaries on the first 5 calls each month. Premium, the lowest tier one person can buy, reads $20 a month in the monthly view and $16 in the annual view. Team reads $19 per user a month monthly and $15 annually, with a two-user minimum; Business $34 and $25. The page loads in the annual view. Both views read on 2026-10-11.
Outbound links may be affiliate links and can earn us a commission: they never touch a score, and the order on this page is the composite order, computed at build time from the cards themselves.
The score cards in full
What each card measured and what it did not: every number dated, sourced and reproducible.
Sends a bot named Fred into Zoom, Google Meet and Teams calls to record, transcribe and summarise them, billed per seat, with a documented API and exports you can edit.
Free tier: 'Free forever', with 'Unlimited transcription*', '400 mins of storage/team' and '20 AI credits'. Paid tiers read $18 (Pro) and $29 (Business) per seat a month in the monthly view and $10 and $19 per seat a month in the annual view, which is the state the page loads in. Enterprise is $39 per seat a month and sold annually only. Both views read on 2026-09-22.
Free plan: 'Free forever', with 'Unlimited recordings + transcriptions'; advanced summaries on the first 5 calls each month. Premium, the lowest tier one person can buy, reads $20 a month in the monthly view and $16 in the annual view. Team reads $19 per user a month monthly and $15 annually, with a two-user minimum; Business $34 and $25. The page loads in the annual view. Both views read on 2026-10-11.
Read the dimensions, not the composite
Price the seat, not the plan. Meeting tools bill per user, and adding a teammate to a workspace usually adds a paid seat, so a team's real cost is the monthly per-seat figure multiplied by everyone who needs to read the notes, not the per-month number a pricing page shows in its annual view.
Then read the storage and credit lines printed under the word unlimited. Transcription can be uncapped while the minutes a plan keeps, or the AI credits that run summaries and follow-up questions, are not.
And read who the terms make responsible for consent. A note taker records everyone on the call. Where a vendor's terms say the host must make sure each participant agrees to being recorded, that duty sits with the buyer under the contract, and the card quotes the clause with its section and the date it was read.
Where these numbers come from
Capability is sourced from public output-quality arenas (blind pairwise votes) or from a published accuracy study where no arena covers the category, usability from review-crowd aggregates weighted by sample size, value from verified pricing, and commercial terms from clause positions read off the vendor’s own legal documents on a stated date. Where none of those exists for a tool, the row carries a named criterion instead: a different measurement, taken from saved sources under its own rubric, and the card prints that criterion’s name and the question it answers in place of the axis heading, so the row is never read as the axis it could not fill. Full detail: methodology.
Cards on this page are ordered by composite, highest first, computed from the scores at build time rather than placed by hand. A tie is published as a tie, and a card with too little measured to carry a composite takes no position.
THE CAPABILITY COLUMN IS LABELLED FOR WHAT IT MEASURES. It publishes the criterion `delivery_envelope_v1`: whether transcripts leave the product in a format you can edit, whether a downloaded file carries the vendor's branding, whether the priced tier can reach the API, and which surfaces the priced tier runs on. It does not say whether a transcript is right, and it is labelled for the question it answers rather than the axis it sits under. When an independent accuracy measurement appears, the cell takes its own name back.
THE USABILITY COLUMN RESTS ON REVIEW PLATFORMS WHOSE SAMPLING CAN BE READ. Where two consumer records each state on their own page how their reviewers were recruited, and state the same thing, they are pooled on that basis. Where neither page states it, the records are pooled by sample size and the card says the recruitment is unknown. Larger app-store samples that state nothing are published in the card's prose with their own counts and dates rather than averaged in.
PRICES ARE MONTHLY AND PER SEAT. A pricing page may load in its annual view and show a per-month figure that needs a year's payment up front; the price on a card is the monthly one, read in both views on the date the card names.