Fellow for iOS: AI Meeting Notes for In-Person Meetings

Fellow’s iOS app does one thing that basically every other AI note-taker still can’t: it works when you leave your desk.

Most meeting transcription tools are built around a pretty specific assumption. You’re on a Zoom call. There’s a bot. The bot joins, records everything, and produces a summary you’ll skim for thirty seconds before closing the tab. That pipeline works fine if your whole professional life happens in browser windows, but a lot of the actually important stuff, the coffee chat where someone mentions they’re leaving their company, the hallway conversation that turns into a product decision, the client lunch where a deal actually gets shaped, none of that runs through a scheduler link.

Fellow’s been around for a while as an AI meeting assistant built around the calendar layer. The Product Hunt listing describes the iOS app as turning “your phone into a powerful AI note taker” for exactly these situations. Record any meeting without a bot or a computer in the loop. Fellow handles the transcription and the AI note-taking automatically. The pitch is pretty direct: if you’ve wanted AI notes during in-person meetings and coffee chats, this is what you’ve been missing.

Aydin Mirzaee, who co-founded Fellow and also co-founded CanvasPop according to his LinkedIn profile, was out promoting the launch himself on launch day. That kind of founder-led push tends to matter for products in this category, where credibility about privacy and reliability is basically everything.

The bot problem, explained in two sentences

Every tool in this space has a version of the same awkward moment. The meeting bot joins the call, a notification pops up that says “FellowAI is recording this meeting,” and then three people on the call make the same joke about being recorded.

In-person, the problem is different. There’s no calendar invite with a video link, so there’s nowhere to inject a bot even if you wanted one. You’re just a person sitting across from another person, and if you want a record of what was said, you either take notes yourself (unreliable, distracting) or you pull out your phone and hit record (which has historically produced an unstructured audio file that you then have to transcribe manually, which no one does). Fellow’s iOS app essentially automates that second path. Hit record, have the conversation, get a transcript and action items automatically afterward.

The app also lets you prep before the meeting. There’s an “Ask Fellow” feature for pre-meeting context, you can edit agendas, and you can set recording preferences directly from your calendar. After the meeting, you can replay audio and video, read through transcripts, review the AI-generated notes, and handle action items by either accepting suggestions Fellow generates or creating your own. That’s a reasonable full loop.

Where this sits in a crowded field

The AI meeting notes category is genuinely packed right now. Otter, Fireflies, Read AI, Notion AI, and a bunch of others all compete for this space. Most of them have mobile apps. The difference Fellow is playing up is the in-person-first angle and the no-bot positioning.

Fellow’s own blog (which they’ve been using aggressively for SEO against competitors like Read AI and Google Meet note-takers) frames the product as supporting both “with or without visible bots, including support for in-person” scenarios. That’s the differentiating sentence. Most tools optimize for the Zoom/Teams/Meet recording pipeline and treat in-person as an afterthought.

The privacy angle also matters a lot here. “Most secure AI meeting assistant” was the language Mirzaee used in his LinkedIn post about the launch. For enterprise buyers, that framing carries real weight. A lot of organizations have real hesitation about bots joining calls and sending audio to third-party servers. A mobile-native recording flow where you control what gets captured and when is a different threat model than a bot that has persistent access to your calendar and auto-joins meetings.

The IAPP’s guidance on AI in meetings and general enterprise software procurement trends both point toward increased scrutiny of exactly this kind of ambient recording tooling. Teams that haven’t fully thought through their policies on AI note-takers are getting pressure from legal and security to figure it out. Fellow’s “most secure” positioning is smart, even if I’d want to see the specific technical claims behind it before repeating them as fact.

The iOS-only question

Launching as an iOS-specific product in 2026 is a deliberate choice. Android gets a substantial chunk of the global smartphone market, and depending on what industry you’re in, it’s not a small percentage of your users. The Fellow team has obviously made a call here, and the likely reasoning is that iOS users tend to skew toward the enterprise and premium segments that are the target buyer for this product.

That said, if you’re building for in-person meetings, which happen everywhere, limiting to one mobile platform creates a real gap. The person who attends the same meeting with an Android device either doesn’t use the tool at all, or someone else on the team runs recording from their iPhone. Neither is great.

The product did well when it launched, ranking fifth across all products on launch day and picking up 215 votes with active comments from people working through real use cases. That’s a decent signal that the core idea resonates with the people most likely to try a new productivity tool.

The deeper workflow question

The thing I keep coming back to with any meeting tool is the action item layer. Transcripts are table stakes at this point. Every tool produces a wall of text that theoretically captures what was said. The harder problem is: what do you actually do with that? Who owns what, and how does it get tracked?

Fellow’s approach here is interesting. The app apparently lets you accept AI-suggested action items or create your own. That’s a step toward genuine task management integration rather than just dumping a transcript into a note you’ll never open again. Whether it actually connects to wherever your team tracks work (Linear, Jira, Notion, whatever the company standardized on three years ago and now everyone hates but nobody wants to migrate away from) is something I’d want to test before making claims about it.

The productivity research literature on meeting effectiveness is pretty consistent: capturing information is the easy part, acting on it is where everything falls apart. A tool that generates good action items but doesn’t connect to execution systems is still better than nothing, but it’s not solving the full problem.

The pre-meeting prep angle

One piece of the Fellow pitch that doesn’t get enough attention: the before-the-meeting layer. “Ask Fellow” for pre-meeting context, editing agendas, setting recording preferences from your calendar. This is a more complete meeting workflow than what most single-feature transcription tools offer.

If you’re going into a client meeting and you want to review the last three months of interaction history, pull up relevant notes, and set an agenda, that’s a different product than “just hit record and we’ll give you a transcript.” It’s genuinely closer to how you’d want a smart assistant to function. The question is execution quality, which I can’t assess from the app description alone, but the design intent is in the right direction.

There’s also something worth saying about the calendar-native approach. Fellow reportedly works off your existing calendar, which means it has context about who you’re meeting, when, and potentially what the stated purpose is. That context makes AI suggestions about agendas and prep much more useful than a blank-slate approach. A general-purpose AI assistant has to ask you a lot of questions to understand what meeting you’re preparing for. Fellow, if it’s doing this right, already knows.

Mirzaee told LinkedIn followers that the in-person meeting scenario was specifically what motivated the iOS push, which makes sense given that Fellow already had the bot-based remote meeting coverage built out. The iOS app reads like filling in the missing third of the meeting surface area, not like a pivot.

The product is free to try, which removes a lot of the friction for someone who wants to test it on a few real meetings before making any commitment. For anyone who spends meaningful time in rooms rather than Zoom windows, that’s probably the right starting point.