Best AI Notetaker for In-Person Meetings (2026, Ranked and Compared)

The best AI notetaker for an in-person meeting depends on the room, not the brand, because in person there is no meeting link for a bot to join, so capture comes from a device microphone instead of a platform stream. For a quiet office, a 1:1, or a sales meeting where you carry a laptop, a botless app like Scribbl captures the conversation, transcribes and summarizes it, and connects the meeting to ChatGPT or Claude so the follow-up gets written, with a free plan and no credit card. For a large, noisy conference table with many speakers, a dedicated multi-microphone hardware recorder separates voices better than a single laptop mic. For solo live note-taking on a Mac, Granola is a strong botless desktop option.

Here is the quick answer, then the full breakdown. The ranked picks, an at-a-glance comparison table, a meeting-type decision matrix, a step-by-step recording walkthrough with mic placement, a hybrid-meeting note, a buying guide, and a consent section are all below. Last updated June 2026.

Scribbl

Quiet office, 1:1, or sales meeting where you carry a laptop, botless, writes the follow-up

Multi-mic

Noisy boardroom with many speakers, a dedicated hardware recorder separates voices best

Granola

Solo live note-taking on a Mac, botless desktop capture

What is an AI notetaker for in-person meetings?

An AI notetaker for in-person meetings captures the audio of a face-to-face conversation from a device microphone (a laptop, a phone, or a dedicated recorder), then uses AI to turn it into a transcript, a summary, and action items. The output is the same thing you would get from a video-call notetaker. The difference is the source: instead of reading a meeting platform's audio stream, it picks up the air in the room.

That single change ripples through everything. There is no meeting link, so there is nothing for a bot to join. Everyone is on one shared microphone, so separating speakers across a table is harder than on a call where each person has their own channel. Room noise, table size, and distance all matter. And because the recorder does not announce itself the way a platform bot might, disclosure is on you.

The best in-person notetakers do more than transcribe. They identify speakers, work offline or on-device when the room has poor Wi-Fi, and push the result into the tools you already use so the meeting turns into action. Scribbl takes that last step furthest by connecting the meeting to ChatGPT or Claude, so the recap and follow-up get written from what was actually said.

Transcript

Searchable record of who said what in the room

AI notes

Summary and action items, not a wall of text

Follow-up

The notes turn into the next step of work

In-person vs virtual AI note-taking: what actually differs

This is the part the keyword is really about, so it is worth being precise. On a video call, the meeting platform hands the app a clean, per-participant audio stream, and there is a link to attach the recording to. In a room, none of that exists. Here is how the two diverge on the dimensions that change your results.

Virtual call

• Audio source: the meeting platform's per-participant stream
• Speaker ID: easy, each person has their own channel
• Joining: a calendar bot can join the link, or a botless app reads your session
• Internet: usually required to be on the call at all
• Consent: a visible bot may announce itself for you

In-person meeting

• Audio source: one device microphone picking up the room
• Speaker ID: harder, everyone shares one mic across a table
• Joining: no link exists, so no bot, capture is on a device
• Internet: offline or on-device capture is a real advantage
• Consent: nothing announces itself, so disclosure is on you

The five practical consequences:

• No link, no bot. The whole calendar-bot model assumes a meeting URL for a virtual participant to join. In person there is none, which is why apps that lean entirely on a join-the-call bot need a separate recording mode for face-to-face meetings, and why a botless approach translates cleanly.
• One microphone, not many. A video call gives the app separate channels per person. A room gives it one shared microphone, so telling speakers apart across a table is genuinely harder. This is the single biggest reason dedicated multi-mic recorders are pitched for in-person work.
• The environment is in the recording. Background noise, a large echoey boardroom, distance, and people talking over each other all degrade quality. A quiet room and a centered device matter more than any feature.
• Offline capability counts. Plenty of rooms have weak Wi-Fi, and some conversations are sensitive. Tools that can record locally on a device and transcribe later have a real edge in person that simply does not come up on a video call.
• Disclosure shifts to you. On a call, a visible bot can effectively announce the recording. In a room, nothing does, so saying it out loud is the only disclosure that happens.

At-a-glance comparison

The full matrix across the axes that decide an in-person notetaker: how it captures audio in a room, speaker ID across a table, offline support, integrations, action-item capture, and price. Pricing figures are vendor-published June 2026 entry tiers and change often; the in-person capture notes reflect how each tool is commonly positioned for face-to-face use as of June 2026. Confirm specifics on each vendor's site.

Feature	Scribbl	Otter	Plaud	Granola	Fireflies	Jamie	tl;dv
In-person capture method	Botless, laptop or phone mic	Phone app, room mic	Multi-mic hardware device	Desktop mic, local	Mobile app, room mic	On-device app mic	Mobile or device mic
No bot, no link needed
Speaker ID across a table	Single-mic limits apply	Single-mic limits apply	Stronger, multi-mic array	Gaps on multi-person	Single-mic limits apply	Single-mic limits apply	Single-mic limits apply
Offline or on-device capture	Desktop app captures locally	Confirm	Yes, records offline	Local capture, Mac focus	Confirm	On-device focus	Confirm
Summaries + action items
Connects to ChatGPT or Claude		–	–	–	–	–	–
Also captures Meet, Zoom, Teams			Audio import	Local calls		Device audio
Free plan	Yes, no credit card	Yes, minute caps	Limited; hardware cost	Limited; confirm	Yes, with caps	Limited; mostly paid	Yes, with caps
Starting paid (June 2026, confirm)	See pricing	~$10-30/user/mo	Device + plan	~$18-35/user/mo	~$10-19/user/mo	~20-30 EUR/mo	~$18-30/user/mo
Best for	Botless, fast follow-up	Mobile live transcription	Noisy multi-speaker rooms	Solo Mac note-taking	Integrations + search	Privacy-leaning capture	Clips and highlights

Match the tool to the meeting type

The fastest way to choose is to start from the meeting you actually have, not the feature list. Map your most common in-person setting to the capture approach that fits it, then confirm the specifics on each vendor's site.

Feature	Best fit	Why	Watch for
Quiet 1:1 or small office meeting	Botless laptop or phone app (Scribbl)	One centered mic covers close speakers cleanly	Keep the device off soft surfaces
In-person sales or client meeting	Botless app, nothing visible (Scribbl)	No recorder in front of the prospect; fast follow-up	Disclose that you are recording
Large, noisy conference-room workshop	Multi-mic hardware recorder	More mics placed centrally separate distant voices	Extra device to carry, charge, and pay for
Solo live note-taking (Mac)	Botless desktop app (Granola)	Merges your typed notes with the AI summary	Speaker-ID gaps on multi-person audio
Confidential meeting, weak Wi-Fi	On-device or offline capture (Jamie, hardware)	Records locally, transcribes later when online	Confirm what works fully offline
Hybrid: some in-room, some remote	Botless call notetaker (Scribbl)	Capture the platform call; ask in-room folks to lean in	Far-side in-room voices may be faint

What to look for in an in-person AI notetaker (buying guide)

In a room, the feature list matters less than these six things. They are exactly the axes the table above scores.

• Ambient and room-mic audio reliability. Can it cleanly pick up several people a few feet away, not just the person holding the laptop? This is the make-or-break axis in person. A single far-field mic struggles in a big room; a multi-mic device or a centrally placed laptop in a quiet room does well.
• Automatic speaker identification. With everyone on one microphone, accurate speaker labels are harder than on a call. Check how the tool handles multiple voices across a table, and expect any single-mic setup to need clean placement to do well.
• Structured summaries and action items. A transcript is the floor. You want a recap with decisions and owners, so the meeting becomes work, not a wall of text.
• Integrations and where the output goes. Does the result flow into your CRM, your docs, your chat, or an AI you already use, or does it sit in a separate app? Scribbl connects the meeting to ChatGPT or Claude so the follow-up gets drafted from what was said.
• Privacy and consent controls. Because nothing announces itself in a room, you want a tool whose capture you control and whose data terms you have read. See the consent section below.
• Language support and offline capture. If your meetings are not in English, confirm the supported-language count and whether the tool transcribes or translates. If your rooms have weak Wi-Fi or the talk is sensitive, prioritize offline or on-device recording.

How to record an in-person meeting, step by step

Most guides skip this, but the recording itself is where in-person notes succeed or fail. Here is a device-agnostic flow that works whether you use a laptop, a phone, or a dedicated recorder.

1. 1

   Pick the room and the device

   Choose the quietest room you can. For a 1:1 or a small, quiet meeting, the laptop or phone you already carry is enough. For a large or noisy table, plan to use a dedicated multi-mic recorder.

2. 2

   Place the mic centrally, within earshot of everyone

   Set the device in the middle of the table, roughly within 10 to 15 feet of each speaker, on a hard surface. Keep it off cushions and bags that muffle sound and away from fans, vents, and laptop speakers.

3. 3

   Disclose and get consent before you start

   Say out loud that you are using an AI notetaker to capture the meeting, and confirm everyone is comfortable. Nothing in the room announces it for you, so this verbal step is the disclosure.

4. 4

   Start the recording and let it run

   Begin capture in your app or device, confirm it is recording, then put it down and actually listen. The point of an AI notetaker is that you stop writing and pay attention to the people in front of you.

5. 5

   Stop, then let the AI build the notes

   End the recording. The app transcribes the room audio, identifies speakers as best it can on a shared mic, and produces a summary with action items. If you were offline, this happens once you reconnect.

6. 6

   Turn the notes into the follow-up

   Review the summary, then send the recap and write the follow-up. With Scribbl, connect the meeting to ChatGPT or Claude and have the email or CRM note drafted from what was actually said, so the loop is closed before you leave the building.

Below is what a finished capture looks like once the room audio becomes structured notes: a clean summary and action items you can send, not a raw transcript you still have to read.

And the step that separates a notetaker from a transcription tool, asking questions and drafting the follow-up across your meetings inside the AI you already use:

1 2 3

1. 1 Place the device dead center on the table, within ~10-15 ft of every speaker
2. 2 Keep speakers facing in; voices carry toward a centered mic
3. 3 Avoid the corner near the door, vents, and the screen fan, the noise sources

The 7 best AI notetakers for in-person meetings, ranked

The picks are ranked for the most common in-person reality: you bring a device, you want clean notes and a fast follow-up, and you would rather nothing visible record the room. Where a tool is built for a different in-person job (a noisy boardroom, a Mac-only workflow), we say so. Every competitor price or feature should be confirmed on that vendor's own site, since they change.

1. Scribbl, the botless overall pick

Scribbl is our top recommendation for most in-person meetings because it brings the same no-bot principle to the room that makes it work on calls. There is no link to join in person and no participant to introduce; you capture the conversation, and the meeting becomes work. It records, transcribes, and summarizes, then connects the meeting to the AI you already use, ChatGPT or Claude, so the recap and follow-up get written from what was actually said rather than just stored. On video calls it is botless across Google Meet, Zoom, and Microsoft Teams, with nothing joining and nothing announced, and it ships as both a Chrome extension and a desktop app. It is rated 5.0 from 3,000 reviews and used by more than 10,000 organizations, with a free plan that needs no credit card so you can test a real meeting before paying.

For in-person use, the practical fit is a laptop on the table in a quiet office, a 1:1, or a sales meeting. As with any single-microphone setup, speaker separation across a big noisy table is harder than on a call, so placement matters; in a quiet room with the device centered, it captures cleanly. The thing competitors mostly do not do is the after-meeting step: instead of stopping at notes, Scribbl lets you ask questions across your meetings and draft the follow-up inside ChatGPT or Claude.

Pros: botless, no participant to introduce in the room or on the call; connects meetings to ChatGPT or Claude so the work gets done; works across Meet, Zoom, and Teams too; free plan with no credit card; Chrome extension and desktop app; 5.0 from 3,000 reviews. Cons: like any single-mic app, speaker separation in a large, noisy room is harder than with a dedicated multi-mic device; capture is anchored to your own device, which is the point but means you run it. Pricing: free plan, no credit card; paid tiers on the pricing page. Rating: 5.0 from 3,000 reviews, across 10,000+ organizations. Best for: quiet offices, 1:1s, and in-person sales meetings where you want clean notes, no visible recorder, and a follow-up that writes itself. More in our botless AI notetaker explainer and AI notetaker for 1-on-1 meetings guide.

2. A dedicated multi-mic hardware recorder (e.g. Plaud), for noisy rooms

For large conference tables, workshops, and noisy rooms with several speakers, a dedicated pocket recorder with a multi-microphone array is commonly the strongest in-person option. The pitch is straightforward and physically true: more microphones, placed centrally, separate distant voices better than a single laptop mic trying to cover a boardroom. Devices in this category are positioned around broad language support and offline recording, so they can capture without Wi-Fi and transcribe later.

The trade-off is that it is a separate device to carry, charge, and pay for (hardware plus, usually, a transcription plan), and it lives outside the apps where your calls already happen. Verify current device price, plan cost, and any accuracy or language claims on the vendor's own site, since these are exactly the figures that move.

Pros: multi-mic array improves speaker separation in big or noisy rooms; offline recording; often broad language support. Cons: extra hardware to buy and carry; separate from your call workflow; verify accuracy and language claims on the vendor's site. Pricing: device cost plus a transcription plan; confirm on the vendor's site. Best for: noisy, multi-speaker conference rooms where a single laptop mic cannot cleanly separate voices.

3. Granola, for solo live note-taking on a Mac

Granola is a botless desktop app, Mac-focused, that captures audio locally while you optionally type your own notes, then merges your notes with the AI summary. For in-person use it shines as a solo, live note-taking companion: you sit in the meeting with your laptop, jot the occasional line, and it fills in the rest. Because capture is local, nothing joins and the audio stays on your machine until processed.

The known limitation is speaker identification on multi-person, single-microphone audio, where gaps have been reported, which matters more in a busy room than in a quiet 1:1. Confirm current pricing and platform coverage on its site.

Pros: botless and local; merges your typed notes with the AI summary; strong solo desktop experience. Cons: Mac-focused; speaker-ID gaps on multi-person audio; confirm pricing on site. Pricing: limited free, then roughly $18-35/user/mo (June 2026, confirm). Best for: solo professionals taking live notes on a Mac in 1:1s and small meetings. See our Granola alternative comparison.

4. Otter, for mobile live transcription in the room

Otter offers a mobile app that can record and transcribe an in-person conversation live, which makes it a familiar choice for capturing a meeting from your phone on the table. The live transcription experience and mobile apps are its long-standing strengths. On video calls it defaults to a join-the-call bot, but the phone-recording mode is what applies in person.

As with any single-microphone phone capture, expect speaker separation and accuracy to depend heavily on placement and room noise, and confirm current minute caps and pricing on its site. Note that Otter's auto-joining bot drew consumer-privacy litigation in 2025 on the virtual side, a reminder to choose your capture method on purpose and disclose in person.

Pros: strong mobile app and live transcription; familiar; free tier available. Cons: virtual mode defaults to a visible bot; single-phone-mic limits in big rooms; minute caps on free; confirm pricing. Pricing: free with minute caps, then roughly $10-30/user/mo (June 2026, confirm). Best for: capturing in-person meetings live from a phone. See our Otter alternative comparison.

5. Fireflies, for integrations and team search

Fireflies is best known as an integration-heavy, bot-based notetaker for virtual calls, with broad CRM connections and strong cross-meeting search. For in-person use it offers a mobile recording mode, so a team already standardized on Fireflies for its CRM and search can keep face-to-face meetings in the same library rather than adopting a second tool.

The in-room capture is still single-microphone, with the usual placement and crosstalk caveats, and the product's center of gravity is the virtual bot. Confirm current mobile-capture behavior, CRM integrations, and pricing on its site.

Pros: very broad integrations and deep cross-meeting search; mobile capture keeps in-person notes in the same library. Cons: built around a virtual bot; single-mic limits in person; confirm in-person capture details and pricing on site. Pricing: free with caps, then roughly $10-19/user/mo (June 2026, confirm). Best for: teams that want one searchable library and deep CRM integrations across virtual and in-person meetings. See our Fireflies alternative comparison.

6. Jamie, for privacy-leaning on-device capture

Jamie is positioned as a privacy-leaning, on-device-focused notetaker without a bot, with broad language support. The on-device angle is genuinely relevant in person: for sensitive face-to-face conversations or rooms with poor connectivity, capturing on your own device before processing is a real advantage over a tool that needs a live cloud connection.

It is mostly paid with a limited free allowance and is priced in euros, and as with any single-mic app, large noisy rooms remain a challenge. Confirm current pricing, language count, and data terms on its site.

Pros: botless and privacy-leaning; on-device focus suits sensitive or low-connectivity rooms; broad language support. Cons: mostly paid, limited free; euro pricing; single-mic limits in big rooms; confirm on site. Pricing: limited free, then roughly 20-30 EUR/user/mo (June 2026, confirm). Best for: privacy-conscious professionals who want on-device, no-bot capture of confidential in-person meetings.

7. tl;dv, for clips and highlight sharing

tl;dv is best known for turning meetings into shareable clips and highlights, a useful pattern when you want to circulate the key moment of an in-person meeting rather than the whole transcript. It supports device and mobile capture for face-to-face use, and the clipping and sharing workflow is its real strength.

Like the others here, in-room capture is single-microphone, and clipping matters more for sharing than for raw accuracy. Confirm current in-person capture details, free-tier limits, and pricing on its site.

Pros: strong clipping and highlight-sharing workflow; device and mobile capture; free tier available. Cons: built primarily for virtual calls; single-mic limits in person; confirm in-person details and pricing on site. Pricing: free with caps, then roughly $18-30/user/mo (June 2026, confirm). Best for: sharing standout moments and clips from in-person meetings. See our tl;dv alternative comparison.

The bot vs botless choice, in a room

On a video call, the bot debate is about a visible participant joining the link. In a room, the bot model is not even on the table, because there is no link, and that is exactly why a botless approach is the natural fit for in-person work. The principle that makes Scribbl quiet on a client call (nothing joins, nothing is announced, no email goes out) is the same principle that fits a face-to-face meeting: the focus stays on the people, not the tool.

A bot-first notetaker, used in person

• Built for: joining a meeting link as a visible participant
• In a room: needs a separate mobile or device recording mode
• Workflow: in-person notes can sit apart from the bot-driven library
• Examples: Otter, Fireflies, tl;dv add a phone or device mode

Scribbl (botless everywhere)

• Built for: capturing without anything joining, by design
• In a room: the same no-bot principle, capture on your device
• Workflow: connects the meeting to ChatGPT or Claude for the follow-up
• Plus: free plan, no credit card, and it works on Meet, Zoom, Teams too

What about hybrid meetings (some people in the room, some remote)?

This is the case the in-person guides skip, and it is increasingly the norm: a few people around a table and a few more dialed in on Google Meet, Zoom, or Microsoft Teams. The good news is that a hybrid meeting still has a meeting link, so you are back on familiar ground. Capture the call the way you normally would, and the remote participants come through on their own clean per-participant streams.

The wrinkle is the people in the room. They share one in-room microphone (whatever device is on the call), so the same single-mic limits apply to the far end of the table: voices that are far from the laptop can come through faint, and speaker separation among the in-room group is harder than for the remote attendees. The fix is the same as pure in-person: center the device, ask people across the table to speak up, and keep the room quiet. A botless tool like Scribbl is a clean fit for hybrid because it captures the platform call across Meet, Zoom, and Teams without a bot announcing itself to anyone, in the room or remote, and then connects the meeting to ChatGPT or Claude so one follow-up covers the whole group. If your hybrid rooms are large and the in-room voices keep coming through faint, a centrally placed conference mic, or a dedicated multi-mic recorder running alongside the call, closes the gap.

A note on consent for in-person recording

Recording an in-person conversation is still recording, and the law treats it the same as a call. Consent rules vary by country and, in the US, by state, with some places allowing one-party consent and others requiring all-party consent, and they apply to face-to-face conversations as well as phone and video calls. The difference in person is that nothing announces the recording the way a meeting-platform bot sometimes does, so the disclosure is entirely on you.

The safe default is simple: at the start of the meeting, say out loud that you are using an AI notetaker to capture the conversation, and confirm everyone is comfortable before you begin. It takes a few seconds and keeps things straightforward later. This is general information, not legal advice; check the rules for your jurisdiction and your company policy before recording sensitive meetings. For a deeper read, see do you have to tell people you are recording and is it legal to record a meeting.

5.0

rating from 3,000 reviews

10,000+

organizations using Scribbl

0

bots in your meeting, in person or on a call

FAQ

What is the best AI notetaker for in-person meetings in 2026?

It depends on the room. For a quiet office, a 1:1, or a sales meeting where you bring a laptop, a botless app like Scribbl captures the conversation and connects it to ChatGPT or Claude so the follow-up gets written, with a free plan and no credit card. For a noisy conference table or a multi-speaker workshop, a dedicated multi-mic hardware recorder (commonly cited examples include Plaud and similar pocket devices) separates voices better than a single laptop mic. For solo live note-taking on a Mac, Granola is a strong botless desktop option. Pick on your most common in-person setting (room size, number of speakers, and whether you carry a device) rather than the longest feature list, and confirm current pricing on each vendor's site.

How is capturing an in-person meeting different from a video call?

On a video call, the platform produces a clean per-participant audio stream and a bot-based or botless app reads that stream, so speaker labels are easy and there is a link to attach to. In a room there is no platform stream and no link, so capture comes from a device microphone (laptop, phone, or a dedicated recorder) picking up the air in the room. That means one shared microphone instead of separate channels, harder speaker separation across a table, more sensitivity to room noise and distance, and a stronger need for offline or on-device recording. The calendar-bot model simply does not apply in person, because there is nothing for a bot to join.

Do I need consent to record an in-person meeting?

Yes, treat it the same as recording a call. Recording laws vary by country and, in the US, by state, with some places allowing one-party consent and others requiring all-party consent, and they apply to face-to-face conversations as well as phone and video calls. Because an in-person recorder does not announce itself the way a meeting-platform bot might, the disclosure is entirely on you. The safe default is to say at the start that you are using an AI notetaker to capture the conversation and confirm everyone is comfortable before you begin. This is general information, not legal advice; check the rules for your jurisdiction and your company policy before recording sensitive meetings.

Does a bot join an in-person meeting?

No. The bot model only works when there is a meeting link for a virtual participant to join, and an in-person meeting has no link. Capture instead comes from a device microphone in the room. Some apps that use a calendar bot on video calls offer a separate mobile or desktop recording mode for in-person use; a botless app like Scribbl uses the same no-bot principle everywhere, which is also why it never puts a visible recorder in front of clients on a call.

How accurate are AI notetakers in a conference room?

Accuracy in a room depends mostly on the microphone and the distance, not the brand. Close, clean audio from one or two nearby speakers transcribes nearly as well as a video call. Accuracy drops with crosstalk, a large table, distance beyond a few feet, hard echoey rooms, and a single far-field laptop mic trying to cover everyone. This is the main reason dedicated multi-microphone recorders are pitched for in-person use: more mics, placed centrally, separate voices better than one laptop mic across a boardroom. For any tool, the honest test is to record your own worst room (a busy table with crosstalk) and read the transcript before you commit.

Can I record an in-person meeting without internet?

Sometimes, and it depends on the tool. Dedicated hardware recorders and some on-device or desktop apps can record audio locally with no connection and then transcribe later once you are online, which is useful in rooms with poor Wi-Fi or for confidential offline capture. Many cloud notetakers need a connection to transcribe in real time. If offline capture matters to you, confirm on each vendor's site whether recording, transcription, or both require internet, since the answer varies by tool and tier.

What is an AI notetaker for in-person meetings?

It is a tool that captures the audio of a face-to-face meeting from a device microphone (a laptop, phone, or dedicated recorder) and uses AI to turn it into a transcript, a summary, and action items, the same output you would get from a video-call notetaker, just sourced from the room instead of a meeting platform's stream. The best ones also identify speakers, work offline or on-device, and push the result into the tools you already use so the follow-up gets done, not just stored.

Which AI notetaker is best for in-person sales meetings?

For in-person sales meetings, where you usually bring a laptop and want a fast, clean follow-up, a botless app like Scribbl is a strong pick: nothing visible records the prospect, and it connects the meeting to ChatGPT or Claude so the recap and CRM note get written from what was actually said. If your sales meetings are large, noisy, multi-person rooms, a dedicated multi-mic recorder can improve speaker separation. Either way, disclose that you are recording, since face-to-face conversations carry the same consent rules as calls. See Scribbl for sales for how the follow-up plays out.

Can I use a phone or laptop as an in-person AI notetaker, or do I need special hardware?

For most quiet rooms, 1:1s, and small meetings, a phone or laptop microphone placed near the speakers is enough, and it is what most people already carry. Dedicated hardware (a pocket recorder with a multi-microphone array) earns its keep in larger rooms, around big conference tables, and in noisy environments where one built-in mic cannot cleanly separate several distant voices. Start with the device you have, record one real meeting, and only buy hardware if the transcript shows the room is too big or noisy for a single mic.

How do I capture a hybrid meeting where some people are in the room and some are remote?

A hybrid meeting still has a meeting link, so capture the call the way you normally would and the remote attendees come through on their own clean per-participant streams. The catch is the in-room group: they share whatever single microphone is on the call, so far-side voices can come through faint and speaker separation is harder than for the remote side. Center the device, ask people across the table to speak up, and keep the room quiet. A botless tool like Scribbl fits hybrid well because it captures the call across Meet, Zoom, and Teams without a bot announcing itself to anyone in the room or remote, then connects the meeting to ChatGPT or Claude so one follow-up covers the whole group. If in-room voices stay faint in a big room, add a centrally placed conference mic or run a dedicated multi-mic recorder alongside the call.