Is there a free AI meeting notetaker?
Yes, several — but "free" means four different things in this category, and picking the wrong one wastes a month. (1) Already paid for: if your company has a paid Zoom, Teams, Google Workspace, or Webex plan, some AI meeting features are usually included at no extra cost — the cheapest option, but it only works on that platform, and on a call another company hosts you don't control whether it's on. (2) Free tiers of commercial notetakers: genuinely usable, and Fathom's is among the most generous, but they're metered — typically by meetings or minutes per month, with history and AI features capped before raw transcription is. (3) Free trials: not free, just deferred. (4) Open-source models like Whisper run on your own machine: free of charge and the only option where nothing leaves your computer, but it costs setup time and a decent machine, and what you get is a transcript, not a summary. Canary's free tier is 5 meetings a month with the full live, bot-free rolling summary on macOS, Windows, and Linux; Pro is $15/month for unlimited. The honest rule: no notetaker is costless — you pay in metered meetings, in your meeting data, or in your own time.
Last updated August 16, 2026
Yes — but the useful answer isn’t a list of tools, it’s a warning that “free” means four different things in this category, and they cost you in four different currencies. People download a free notetaker, hit a wall three weeks later, and conclude the category doesn’t work for them. Usually they just picked the wrong kind of free.
The four kinds of free
| Kind | What it is | What it really costs | Best when |
|---|---|---|---|
| Already paid for | AI features included with your company’s paid Zoom, Teams, Workspace, or Webex plan | Nothing extra — but it’s one platform only, and the host’s plan and admin settings decide whether it exists on a given call | Your whole week happens on one platform you host |
| Free tier | The permanently free plan of a commercial notetaker | Metered: a monthly cap on meetings or minutes, plus limits on history and AI features | You’re in a handful of calls that matter, or evaluating before you buy |
| Free trial | Full product, fixed window | Deferred, not avoided — the bill arrives on day 15 | You’re testing whether the paid plan is worth it |
| Open source, self-run | A model like Whisper running on your own machine | Setup, maintenance, and hardware — and you get a transcript, not notes | Privacy is the requirement and you’re comfortable in a terminal |
Everything below is just those four rows with the fine print filled in.
The free option most people already have
Before installing anything, check what your employer already bought. If the company pays for Zoom, Microsoft 365, Google Workspace, or Webex, platform-native meeting AI is often bundled into the plan at no additional charge, and it comes with two advantages nothing else can match: it’s already covered by the security review IT did, and it’s inside the meeting pipeline, so it gets speaker identity from the participant roster instead of inferring it from audio.
Two catches, both structural. First, it stops at the platform’s border — a Zoom assistant knows nothing about your Teams call, so a multi-platform week yields a patchwork. Second, and more importantly, it’s a feature of the meeting, not of you. On a call a customer or client hosts, their plan and their admin settings decide whether any AI runs, and your own subscription doesn’t follow you in. That’s the whole problem with the free-because-included option when you’re not the host — and cross-company calls are disproportionately the ones you most need notes from. Availability and packaging also shift often and vary by plan, edition, and region, so check your own account rather than a comparison table.
What free tiers actually cap
Vendors change limits frequently and none of the specifics below are worth memorizing, but the shapes of the caps are stable across the category:
- Meetings or minutes per month. The headline limit, and the one that decides whether a free tier is a plan or an evaluation.
- History and search. Many free tiers keep recent meetings and quietly age out older ones, which matters most for the exact use case — finding what someone committed to two months ago — that made you want a notetaker.
- The AI layer before the raw layer. This is the pattern to watch: transcription is cheap to serve, so free tiers tend to be more generous with transcripts than with summaries, action items, templates, and chat-with-your-meeting features. You can end up with plenty of words and none of the parts that save you time.
- Export and integrations. Pushing notes into Notion, a CRM, or Slack is a common paid-plan line.
- Seats. Free is usually per-person; the moment a team wants shared notes it becomes a per-seat conversation.
Credit where it’s due: some free tiers are genuinely good. Fathom’s is among the most generous in the category, and if you’re comfortable with a bot joining your calls and reading notes afterward, it’s a reasonable place to start — the full comparison covers the trade-offs.
The currency you don’t see on the pricing page
A free hosted notetaker is still a product with costs behind it, and the honest question is what’s funding it. Usually it’s a funnel — the free tier exists to sell the paid one, which is straightforward and fine. Sometimes the economics lean on the data instead. Either way, on essentially every hosted tool, free or paid, your meeting audio or transcript is processed in someone’s cloud; that’s as true of bot-free tools as of bot-based ones, because the model doing the transcribing usually isn’t running on your laptop. What varies is retention, whether your content is used for training, and who the named subprocessors are. Are AI meeting notetakers safe? covers what to look for; the short version is that the answer should be in the privacy policy, and if it isn’t, that’s the finding.
The second invisible cost is your own time. A free tool that produces a clean document you read forty minutes later still leaves you doing the reading — and if the reason you wanted a notetaker was that you’re already behind, that’s the cost that actually hurts.
Rolling your own
If “free” means “no subscription, ever” and privacy is the driver, running an open-source model locally is a real answer rather than a theoretical one. Whisper-family models are open source, community desktop wrappers make them approachable, and nothing leaves your machine — which no hosted product can claim.
Be clear-eyed about what you get. The pipeline every notetaker runs has four stages — capture, transcribe, separate speakers, summarize — and a local model hands you the middle one. Getting audio in requires solving capture yourself (a loopback or monitor device, per OS), speaker separation is a separate model, and summarization means wiring up an LLM, at which point you’re either paying an API bill or running a local model too. It’s a good weekend project and a poor Tuesday-morning plan. If you’re on Linux, the Linux notetaker answer walks through why capture in particular is the hard part.
Where Canary’s free tier fits
Canary is a real-time, bot-free meeting summarizer. It captures your computer’s system audio (no bot in the call, no plugin) and shows a live, multi-resolution rolling summary — from what’s being said right now to the whole call — so you can catch up the instant your name is called.
The free tier is 5 meetings a month, on macOS, Windows, and Linux, with the rolling summary itself included rather than held back for the paid plan — the live layer is the product, so capping it would make the free tier pointless. Pro is $15/month for unlimited meetings.
The honest framing: 5 meetings is a real week’s worth of the calls that matter, and it’s deliberately enough to find out whether a during-the-call summary changes how your meetings go. It is not a plan for someone in six calls a day. If you’re in that many meetings, treat every free tier in this category — Canary’s included — as an evaluation, and decide based on whether the tool earned its keep in those five calls.
How to choose in one pass
- Everything happens on one platform, and you host it. Use what your company already pays for. Start with the built-in assistant.
- You join other companies’ calls. Platform AI will keep failing you there — you can’t switch on someone else’s tenant. Use a tool that runs on your own machine.
- You want a document to read afterward and don’t mind a bot. A generous bot-based free tier is the cheapest path to good notes.
- You want no bot and no bill. A bot-free tool’s free tier, or a local Whisper setup if you’ll do the work.
- You keep getting caught out mid-call. That’s a timing problem, not a price problem, and no post-meeting tool at any price solves it — see best bot-free meeting notes for how the options compare.
Free doesn’t change the disclosure
One last thing, because it gets skipped precisely on free tools: a notetaker that costs nothing is still a tool capturing a conversation other people are in. The obligation is identical at $0 and $15. Platform-native AI at least announces itself — a banner or an icon appears — so if you’ve moved to a tool running outside the meeting, the announcing is your job. How to tell participants you’re using an AI notetaker has phrasing that takes one sentence, and consent rules vary by region — some require only one participant’s consent, others everyone’s. Default to saying so.
Frequently asked questions
Is there a completely free AI meeting notetaker with no limits?
Realistically, only one you run yourself. Open-source speech models such as Whisper can be run locally through community desktop wrappers with no subscription and no meeting cap, and because the audio never leaves your machine it's also the most private option available. The costs are real but non-monetary: you install and maintain it, you need hardware fast enough to keep up, and it produces a transcript — turning that into structured notes or a live summary is the part you'd be building yourself. Every hosted product with a genuinely unlimited free tier is either an early-stage tool that will introduce limits later, or is monetizing something other than the subscription, which is worth reading the privacy policy to understand.
Do free AI notetakers use my meeting data to train their models?
It varies by vendor, and it's the question worth actually checking rather than assuming — free products have fewer revenue sources than paid ones, so the incentive is different. What to look for in the privacy policy: an explicit statement that your content isn't used for model training, a named list of subprocessors (the cloud speech and language services that actually see the audio), a stated retention period, and whether the default differs between free and paid plans. Note that on essentially every hosted tool, free or paid, your meeting audio or transcript is processed in someone's cloud — that's true of bot-free tools too. The only architecture that avoids it entirely is running a model locally.
Is a free meeting notetaker the same as a bot-free one?
No — they're different words for different things, and the category makes them easy to confuse. "Free" is about price. "Bot-free" is about whether a separate participant joins your call to record it: bot-free tools capture your computer's system audio instead, so nothing appears in the participant list and nothing has to be admitted. Plenty of bot-based tools have free tiers, and plenty of bot-free tools charge. If you want both — no bot and no bill — the practical options are a bot-free tool's free tier or an open-source setup you run yourself.