Launched this week
Pauta
Your private memory for every conversation
78 followers
Your private memory for every conversation
78 followers
Pauta records and transcribes meetings, interviews and voice notes, then turns them into summaries and action items. It works fully offline, keeps your data on your device, and only connects to external AI if you choose.
Free Options
Launch Team / Built With
AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
Hey Product Hunt 👋
I’m Paulo, one of the founders of Foreigners, and together with Fê and Leo we built Pauta.
Pauta started from a simple problem: a huge amount of useful information lives inside conversations — meetings, interviews, voice messages, spontaneous ideas — and most of it disappears as soon as the conversation ends.
We wanted a way to keep that context without turning every conversation into another cloud data source.
So we built Pauta around a simple idea: your conversations should become useful memory, while remaining yours.
Pauta can record and transcribe conversations, generate summaries and action items, and help you revisit what was actually said.
But the part we care deeply about is how it does it.
Your recordings and data stay on your device by default. Pauta can work fully offline. If you want to connect an external AI, that’s your choice — not a requirement for the product to work.
We’re already using it ourselves for meetings, job interviews, voice messages and those moments when an idea shows up and you just want to start talking.
Pauta is available today on macOS and iOS, and we’re currently testing Windows.
We’d love to hear how you’d use something like this — and especially what you think a truly private AI memory should be able to do next.
— Paulo
Foreigners INC.
voice messages are the worst for this, someone explains something important for 4 min and a week later nobody can find it. does it work with telegram voice notes?
Feel free to ask new features and send feedback, we are continually growing the project and new ideas are always welcome!
The local-first default is what makes this interesting — most note-takers nail the summary part and then quietly send everything to the cloud for search and Q&A. Keeping transcription and summaries on-device, with external AI as an opt-in instead of the price of admission, is the harder design choice and the right one. Curious how the local model holds up on longer meetings — does it chunk things pretty well, or is there a length where context starts to get thin?
The product decision that matters is when text leaves the machine. If local Qwen is the default for summaries, what still goes out when someone later asks a question across a whole project?
@niklausgerber Nothing goes out, you continue to only using the local models, the project data is all stored on your own computer.
Thanks. That is the test that usually breaks. Summaries stay local, then project Q&A phones home. Good that yours does not.