Ellis - AI notes for in-person meetings
by•
Ellis is an AI notetaker for in-person meetings. Record your meeting, get a clean transcript with each speaker identified, then ask anything — what was decided, what you missed, how it went. No laptop. No extra hardware. Just your iPhone (or Apple Watch).
Replies
How well does the speaker identification actually hold up when people talk over each other or move around the room, especially in louder spaces like cafes or open offices?
Ellis
@sabanl51707 So I tested cafes and open offices. That is my main testing ground and it works quite well. But give it a try yourself
Octolens
Congrats Robin!! Ellis looks super sleek.
Ellis
@charlotteschmitt appreciate it!
Used it during a quick team sync and the speaker labels were actually right on, which I was not expecting. Being able to just ask "what did we decide" after is the part that sold me.
Ellis
@birol160281 happy to hear it.
the noisy-room questions are all covered, mine's about consent instead - you list therapy and doctor visits as use cases, which means the other person in the room is being fully transcribed and analyzable without necessarily agreeing to that. recording is auto-deleted but the transcript sticks around indefinitely. a lot of US states are two-party consent for recording, does Ellis do anything to prompt "hey, tell the other person" or is that entirely left to the user's judgment
Ellis
@galdayan yes! I'm adding an info message before the start of each recording to remind users of consent. I live in Germany, where it's even more pressing :)
When it comes to therapy in particular, I've found that many therapists actually recommend their clients to take notes/record. But that's a very trusted relationship.
Do you have specific thoughts here?
@galdayan @robingreenwood I think the useful move is making consent part of the recording flow, not just a reminder before it. For sensitive contexts like doctor visits, therapy, caregiving, or interviews, I’d trust it more if each recording began with a simple “I have permission to record this conversation” confirmation and then kept retention/deletion settings visible after transcription. The recording disappearing is good, but the transcript is still the durable memory, so that layer needs to feel just as intentional.
Ellis
@galdayan @eduard_turea like a toggle button that when "on" enables recording to start?
@galdayan @robingreenwood Yes, something in that direction, but I’d make it per recording rather than a general toggle that stays on. Before capture starts, the app could ask for a simple permission confirmation for this conversation, and after transcription keep deletion and retention controls visible on the transcript. That ties the consent moment to the durable note, not only to the audio recording.
How well does it handle meetings with more than like 6 people talking over each other, especially in a noisy cafe setting?
Ellis
@birsen338162233 Give it a try. But talking over each other will be difficult in any situation.
How well does the speaker identification actually work in a noisy room with multiple people talking over each other? Curious how it handles real-world chaos versus a quiet conference room.
Ellis
@boran925180 So real-world chaos is, of course, going to be harder. But typically those are not the most productive meetings nor where you would want to extract valuable information from. Give it a try
the speaker identification worked surprisingly well in a noisy cafe setting, way better than i expected from a phone mic.
Ellis
@aysimag92529 nice!
Everyone solved Zoom notes, nobody solved the conference-room whiteboard session. In-person is the harder and more valuable problem. How does it handle multiple speakers without per-person mics?
Ellis
@medal411 There are multiple steps under the hood.
Voice enrollment, which captures voice attributes
Diarization using assembly AI? And finding a best match to your voice
A UI that lets you easily pick and assign speakers
@robingreenwood Thanks for breaking that down — enrollment capturing voice attributes + diarization, then best-match, makes sense. The speaker-assignment UI is the part most tools get wrong, so smart to make that easy. Congrats on the launch!
Ellis
@medal411 appreciate it!
How does it handle background noise and overlapping voices in a real conference room setup?
Ellis
@ferideayku48864 It uses AssemblyAI to diarize and identify the different speakers. But you should give it a try for your real conference room setup.
The speaker identification was actually solid in my quick test, which I didn't expect from a phone-only setup. Wish I'd had this for the standup meetings I always forget half of.
Ellis
@ykselwil7 happy to hear it