Murmur
A private voice-first notepad for your thoughts
51 followers
A private voice-first notepad for your thoughts
51 followers
Murmur is a private, voice-first memory for your thoughts. Speak instead of typing, and Murmur turns your fleeting thoughts into searchable memories. Later, ask what you were thinking about, rediscover forgotten ideas, or send a thought to your future self with Future Capsules. No feeds. No followers. Just your thoughts, remembered.
Most apps make you capture thoughts and leave you with a pile of recordings.
Murmur is built to help you capture your thoughts by voice, give them structure, and revisit them when you need them.
No need to type everything out or organize every thought in the moment. Just speak, and come back to it later.
Murmur is available on Android. I'm looking for people to try it and share honest feedback.
https://murmurr.subhr.in
the part I'd want to understand before trusting this with real thoughts is the search side, not the capture side. recording a voice note is the easy half, people already do that with a plain voice memo app. the hard part is that when I actually go looking for something later I never remember my own phrasing, I remember the gist. does search here work off the literal transcript or off some semantic layer on top, because "find that thing I said about the vendor contract" only works if it's not doing keyword matching against whatever rambling words I actually used in the moment
@galdayan Exactly, for example, you might record a thought saying, “We should reconsider the pricing before finalizing the deal.” A few weeks later, you might search for “that thing I said about the vendor contract” without remembering your exact words. Ideally, Murmur should help you find that thought based on its meaning, not just matching keywords. That's the kind of experience I'd like to build.
@subhrajyoti_sahoo "ideally" is the word doing all the work there. is that semantic matching live in the app today, or still the direction you're building toward? asking because that's the actual hard engineering problem, matching meaning instead of keywords on short rambly voice transcripts, and I'd want to know if it's shipped before I trust it with the kind of thought I'd never phrase the same way twice
@galdayan Still working on making it accurate. It works in most cases, but it's still failing in a few instances and edge cases.
@subhrajyoti_sahoo fair enough, every search feature says that at this stage. what do the edge cases look like in practice - is it mostly short notes with no context, or notes where you used a word that's semantically close but not the one your brain will search for later?