cd /news/artificial-intelligence/the-ai-notetaker-has-been-invited-to… · home topics artificial-intelligence article
[ARTICLE · art-87299] src=wired.com ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

The AI Notetaker Has Been Invited to All the Meetings

Wispr Flow, the AI voice dictation startup led by CEO Tanay Kothari, launched Notetaker, a feature that transcribes and summarizes meetings in real time, available first on Mac with Windows support planned. The tool connects to calendars, offers a six-hour maximum recording length, and includes a Privacy Mode that excludes dictation data from AI training. Kothari emphasized that users should disclose when using the tool, citing legal and trust considerations.

read4 min views1 publishedAug 5, 2026
The AI Notetaker Has Been Invited to All the Meetings
Image: Wired AI

As Wispr Flow CEO Tanay Kothari chats with me about his company’s new Notetaker feature, my eyes keep bouncing to the corner of my laptop screen where Wispr’s AI tool is transcribing everything we say. It’s a verbatim record of our conversation generated by AI in real time, even down to my “mm-hmm” sounds while I listen to Kothari explaining why the startup, which is known for voice dictation, is dropping something new.

“This is the year of meeting recorders and notetakers,” says Kothari. And it’s hard to argue with that, at least in Silicon Valley. Most tech investors and other experts I’ve chatted with this year now assume whenever they meet up with someone to talk business that the conversation is going to be recorded or transcribed by AI, sometimes without explicit consent. The current option du jour for meeting records is Granola AI, with plenty of other options, like Otter AI, also available to users. So, the timing of Wispr’s expansion from a dictation tool into a meeting resource makes sense.

Wispr’s Notetaker connects to your calendar and can listen along during any meeting. After the conversation is over, Wispr uses generative AI to summarize the conversation and connect it to a chatbot-style tool for parsing its context with follow-up questions. The tool is launching first on Mac laptops, with a Windows expansion expected down the line. If you enable Privacy Mode for your account, then the data from your dictations will not be used by Wispr for AI model training.

In my initial test using Notetaker during a conversation with the CEO, the live transcription appeared accurate as it flashed across the screen, and I didn’t spot any obvious errors in its AI summary generated after our meeting. The “ask anything” tool felt like the feature’s best aspect. I had a quote in mind that I wanted to pull out, but didn’t remember when Kothari said it during our chat. So, I typed in “year of ai notetakers quote,” and the Notetaker tool found the moment that I remembered. When I checked it against my audio recording for accuracy, the AI-transcribed quote matched the audio file.

Users can fiddle with the fine details of Wispr’s Notetaker in the Mac app’s settings. One of the most notable aspects available here is the ability to extend the maximum meetings recording length to a whopping six hours.

“Wispr had a six-minute time limit earlier. People would start it on and off five times just to capture a 30-minute meeting,” says Kothari. “We realized we should stop being stupid. This is what people really want Wispr to do.” Though, if you have the gall to invite me to join a six-hour meeting, just know that I’ll be blocking you on every platform imaginable.

Similar to Granola’s tool for capturing meeting transcripts, Wispr transcribes conversations directly on your device and doesn’t join your meeting as a visible participant. Despite that, Kothari says users shouldn’t secretly use Wispr’s new tool and always disclose when it's transcribing conversations, potentially for legal reasons as well as out of respect for the privacy of the meeting’s participants.

“We need that to become the norm for everybody. Otherwise, you're going to get a lot of lack of trust that starts to happen,” he says. “There are laws that are single-party and multi-party. I think, regardless, it's just a basic trust-setting exercise to just make sure people know.” In San Francisco, where Wispr is based, someone recording a private conversation is legally required to acquire consent from all participants.

Voice dictation capabilities have rapidly improved in recent years with the maturing of underlying AI models, and I often find myself relying on tools like Wispr and its ilk in my daily life. Whether I’m organizing my notes on an academic research paper or searching online for reality TV updates about the current season of Big Brother, I’m constantly blabbing at my devices and expecting crystal clear transcripts, even though I’m prone to mumbling.

What I don’t expect is for anyone to secretly transcribe my errant thoughts during a conversation. Break my trust even once, and it should be socially acceptable to ice you out of my life. And I hope you’re writing that down.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @wispr flow 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/the-ai-notetaker-has…] indexed:0 read:4min 2026-08-05 ·