In this live demo, Christopher Pedregal introduces Granola, an AI‑powered notepad designed for back‑to‑back meetings. Launched in May with a four‑person team, Granola has grown to eight members. The product is a macOS text editor that syncs with the user's calendar, joins Google Meet calls, and captures real‑time transcription by tapping directly into the Mac's microphone and system audio—no external bot is needed. Granola also ingests screenshots and user‑typed notes, then uses AI to "enhance notes" after the meeting, producing polished, context‑rich minutes. The AI enriches notes with participant information pulled from the web and applies persona‑specific templates (e.g., VCs, founders, LPs) to tailor the output. An integrated AI chat lets users query a single meeting's content. Pedregal highlights observed behavior changes, such as users preferring to ask questions rather than reread raw notes. Looking ahead, Granola aims to automate post‑meeting workflows like drafting linear tickets with quoted user statements, while maintaining a human‑in‑the‑loop philosophy that keeps users in control of final decisions. The demo concludes with an invitation for questions and further contact.
Chapters
All right. Thanks for the intro. Granola is an AI notepad for people in back-to-back meetings. Who's tried it here? Whoa. Okay. That definitely was not the answer. That used to not happen. So that's pretty cool. We, let's see what happened. We launched Granola in May. We were a team of 4 at the time. We're now a team of 8. Um, and I'll just show you the product so you see what it's like. So, uh, this is Grenoble over here. Um, it's connected to your calendar, so I can click on a, on an event. And, um, I'm now going to join a Google Meet call with our CTO, Vas, who's hopefully in the back and on there. Hey Vas!
Thank you, how's it going?
Great. You want to tell everybody what Granola is, how it works?
So Granola is a text editor for macOS. So you can write your notes there just like in Apple Notes. But it also transcribes in real time. And when the meeting ends, it will use the context from the call to enhance your notes and make them look great.
Great. Awesome. And how does the transcription work? How does it know what you're saying?
Yeah. So, Renauld runs on your Mac. So it basically taps directly into your microphone and system audio for transcription. So there's no non-bot that joins your call. And it works with any video call in that.
Awesome. All right. Thank you so much, Fas. Cool. Bye. Cool. So that was a quick demo of Granola. You see, I just typed some stuff in there. Something cool that we recently added, let's see if I can do it with one hand, is you can now do screenshots. So it'll take the context from what's spoken, but it'll also take the context from any screenshots or images you might add. And then at the end of the meeting, you can hit enhance notes. And if the demo gods smile upon us today, you'll see some notes being generated. I guess what's interesting here, well, one is like it'll incorporate the notes you write and that gives you as much or as little control as you want. You can take no notes and Granola will do the best it can. You can take lots of notes and it'll basically just try to flesh out your notes and make them better. We've spent a lot of time trying to make it as simple as possible. Like, I think it probably looks very simple and takes a lot of effort to get it to that point. So a big focus for us has been the note quality. And to be able to generate good notes that don't feel generic or that miss the point is actually a really hard task. So to give you a sense of the types of things that we would do is we would, for example, We know that Vas is in this call. We pull information about him from the internet. We try to give the model as much context about who's in the call and what we're gonna be talking about. And then we're very, very opinionated about what makes good notes for a specific type of meeting. Let's see, a good example, like we have a lot of VCs using us. So for example, if you're a VC, if you're an investor and you're meeting with another investor versus a startup founder that's pitching you versus an LP that you're asking for money, Your notes should look completely different. You care about very different things. If you're a startup founder, you're pitching your company, you don't want any information about your company in the notes. You know all that stuff. So there's all these nuances that really, really matter and make a really big difference, but hopefully is invisible to the user. So that's the main product. Something interesting we've observed as more people are using it is written notes like this are great for other people, but most of the time you don't go back and look at them. You'll go back and you'll ask questions. We basically, the AI chat feature we have, like, it's still very basic. You can only speak with one meeting at a time, but we've definitely seen behavior change. Like, that's the main way people go back to look up information. I'm seeing some nodding heads for people who are using it. Cool. And I guess I'll leave some time for questions, but in terms of What we want to do next, meetings are, our view is that we want to build tools that help people think and do work better and meetings are a stepping stone towards that. It's a way for us to earn our right to be in the meeting. All the really cool shit we want to do I think comes afterwards. So it's like something we have a demo for which we haven't launched yet is let's say you do a call with a user and they talk about some bugs, right? It turns out AI is really, really good at drafting a linear ticket based on what the user said and pulling in quotes for them. There are all of these post-meeting workflows that I think are really, really primed for AI. Just so you know about us, our philosophy is that we think it's really silly to try to delegate and automate things completely. We want to keep the human in the loop. We just want to automate all that busy work. We can draft the ticket for you, we can put in the right quotes, but you know what the really key part of that user issue was. So it's like you should really have that final control at the end of the day. We should design tools so that humans stay in control, otherwise I think we're in trouble. Yeah, I guess with that, maybe probably have a few minutes for questions. We're nearby if you ever want to grab coffee, come by, hang out, talk about meetings. We're here, or send me an email. Yeah.
