
Hemory
Keep listening. Searchable memory for your AI agents.
640 followers
Keep listening. Searchable memory for your AI agents.
640 followers
Hemory comes from Hear + Memory. It's not meeting transcription: Hemory listens on the phone or Apple Watch you already own, and keeps every conversation as memory. Your day is auto-split into moments with speaker labels, then settles into a private, searchable memory. Connect it to Claude, Codex, Cursor, or any agent over MCP, and your AI finally has real context to revisit what you heard and build on it.








Launch Team

CloudflareDeploy serverless functions globally in seconds
Promoted
my memory is so bad at this point I'll try literally anything, this included.
@sy_jin_3 Ha, you're not alone. I built Hemory because I had the same problem. Every day there's so much to remember, but only a small part of it really matters compared to the creative, big-picture thinking we actually want to do. Attention is limited, so our idea is simple: let the machine handle the tedious remembering (it's great at that), and keep your attention for the memories and thinking that are worth it. Would love to hear how it works for you!
@sy_jin_3 Ha, same, that's exactly why I built it. Give it a day and let it do the remembering for you.
Listen for a day → connect your AI → ask anything.
@sy_jin_3 Ha, you’re not alone—I built Hemory because I kept running into the same problem. There’s so much to keep track of every day, and I’d rather spend that attention on ideas than trying to remember every detail. That’s the idea behind Hemory: let it help with the remembering so you have more room to think. Would love to hear how it works for you!
@sy_jin_3 Haha, maybe bad memory is just part of being human. Good thing we’re creative enough to invent AI—and use it to build tools like Hemory to help our forgetful brains out lol.
we record calls at Dial too, always disclosed to the person on the other end since it's a business line. an always-on phone mic picking up everyone around you, including people who never agreed to it, seems like the harder problem to solve than the memory/search part. how does Hemory handle consent when it's the other person in the room or on the call, not the account owner, who's being recorded
@galdayan Thanks for raising this, and you're right that it's the harder problem. I won't claim we've solved it. What we do today: audio is discarded by default, so we don't keep recordings of anyone, and on iPhone and Apple Watch the system recording indicator stays visible while Hemory is listening. You can stop it with one tap and delete anything completely. Recording laws vary by region, and users are responsible for following the ones where they are. Beyond that it comes down to disclosing, the same way you do at Dial. If you've learned anything on this at Dial, I'd love to hear it.
@wang_yingqi appreciate the honest answer, discard-by-default plus a visible indicator is the right baseline. one thing I'm still curious about - once a conversation turns into searchable memory, does a third party's voice or name ever get flagged or redacted, or does it just sit in the account owner's memory graph like any other note
@galdayan You’re right—disclosure on a business call is a much clearer setup. Giving the account owner recording controls doesn’t establish consent from everyone around them.
Hemory currently relies on the person recording to tell others and get their agreement before starting, and to stop if someone objects. That leaves a real gap with continuous listening when someone new joins a conversation. We don’t have an automatic way to establish their consent, and I wouldn’t claim that stopping or deleting afterward solves that.
@shawn_xiao1902 that's a fair line to hold, and honestly more trustworthy than claiming a fix you don't have. the "new person joins mid-conversation" gap is probably the sharpest edge case here - even a chime or visible badge when a new voice is detected would at least surface the moment where consent needs to happen, even if the app itself can't enforce it. curious if that's on your roadmap or if it's more of a "can't solve it, so we're not pretending to" situation.
this looks pretty good, gonna give it a try and see how it goes for me.
@mi_leeli Thanks for giving it a try! Let it listen for a full day, then connect it to your AI over MCP and ask what you talked about. That's where it really clicks.
@mi_leeli Thanks for giving it a try! Let us know how it goes—and if anything feels confusing or gets in your way, we’d love to hear that too.
I often forget important things during meetings, and this software helps me record and organize key points from meetings—it's very useful.
@wayne66 Thank you so much, glad it's helping!
Listen for a day → connect your AI → ask anything.
@wayne66 Glad it’s helping! It’s easy to miss a detail when you’re trying to follow the conversation and take notes at the same time. Hope it gives you a little more room to focus on the discussion.
@wayne66 I’ve got an even bigger problem—I sometimes zone out during meetings. Hemory has been a lifesaver for me!
@yinn1014 Thank you! Hemory automatically splits your day into separate conversations, so nothing gets scattered. Then you can ask your AI to pull together everything from this week's meetings on any topic.
@yinn1014 Thanks! Several meetings in a day can really blur together. Being able to search for a detail you remember without first figuring out which meeting it came from is a big part of why we built Hemory. Would love to hear how it works for you!
@yinn1014 You can chat with your AI to pull it all together and connect the dots between meetings—even spot little connections you might’ve missed.
Okay, this is actually pretty cool . Being able to remember all those random conversations could be really useful
@carlvert Glad it clicked for you! I think random conversations are one of the most underrated sources of good stuff. A restaurant tip, a half-formed idea over lunch, a small detail about someone you care about. None of it feels important in the moment, so we never write it down, and it slips away. Hemory listens in the background and keeps those moments searchable. Later you can just ask things like "what was that book Sam recommended last week?" and get it back. Would love to hear what you rediscover!
@carlvert Thank you! The random ones are often the most valuable, and those are exactly the ones you'd never think to record.
Listen for a day → connect your AI → ask anything.
@carlvert Thanks! Sometimes a passing comment only becomes useful days later—right when you can’t quite remember what was said. That’s one of the moments we built Hemory for.
I no longer worry about missing key discussions by forgetting to hit record. Being able to query AI against my full day‑long conversation memory really boosts my work productivity.
@tomzoe This is so great to hear! Querying your whole day is one of the core use cases we built Hemory around. If you haven't yet, connect it to your own agent through MCP. Your agent can then combine your memory with its own reasoning and tools, and that's where the answers get really good. Then, every week or month, ask it to look back: "What themes kept coming up in my meetings this month?" or "What did I commit to that I haven't followed up on?" Patterns you'd never notice day to day start to pop out. Would love to hear what you discover!
@tomzoe I no longer worry about missing key discussions by forgetting to hit record. Being able to query AI against my full day‑long conversation memory really boosts my work productivity.
@tomzoe Exactly! No matter how many days have passed, those memories remain as fresh and reliable as they were in the moment.