01On the street
Talking to people in town
- When
- Sat, Sep 12, 2026, 16:10
- Where
- Shopping streetRecorded by GPS
- Who
HiranoVoiceprint matchSpeaker 2 · Local residentNew voice

Beta now available on all four platforms (until December 31, 2026)
Washa hears exactly who spoke,when, and where.
Speaker diarization in up to 10 languages,
entirely on your own device.
Speaker 1
Hirano
Tomoyasu Hirano, Washa developer
HiranoVoiceprint match
Meetings, planning sessions, casual chats.Every day,we exchange so many words.And what we lose most easily
is our voices.Unlike written words,spoken words carry the speaker’s pauses, their breath, the air in the room —each person’s words, passed back and forth.Most of them vanish without ever being recorded.
Speaker 2
Kamii
Tsubasa Kawakami, developer
KamiiVoiceprint match
A voice carries what a person thinks,and the reasons behind what they decided.Because none of that was kept,the same conversations have been repeated again and again.That’s why
we built Washa.
Speaker 3
Komomo
Momoko Tohara, agent design
KomomoVoiceprint match
It can keep listening, without ever stopping.So at last, capturing every voice,all the way through, has become possible.Meetings, small talk, a word that slips out.People’s voices no longer slip away.
Speaker 4
Gaku
Gaku Noguchi, designer
GakuVoiceprint match
And then it struck us.If the words we speakstay tied to the person who said them,we’ll want to go out and talk with more people.On the street. In shops. With friends.With people we’ve never met.To listen closely, take in their words,and pass them on to the next person.
Where to use it
We leave the round table
and go out to listen.
To people on the street. To discussions in class. To the meeting table.
And to the friends closest to us.
When, where, who said what,
and to whom.
Whose voice, by voiceprint. Where, by GPS.
All kept together on a tag for each moment.
01On the street
HiranoVoiceprint matchSpeaker 2 · Local residentNew voice
02In class
Nakamura · StudentVoiceprint matchSpeaker 2 · StudentNew voiceSpeaker 3 · StudentNew voiceSpeaker 4 · StudentNew voiceSpeaker 5 · StudentNew voice
03In a meeting
Kawamura · Section managerVoiceprint matchSpeaker 2 · Client’s directorNew voice
Nakano · Team leadVoiceprint matchSpeaker 4 · ClientNew voice
04With the team
HiranoVoiceprint match
KamiiVoiceprint match
KomomoVoiceprint match
GakuVoiceprint matchWhat is Washa
AI transcription with speaker diarization,
Transcribe an hour-long meeting in one go, entirely on your own device.
No cap on monthly hours or on the length of a single recording. Hour-long meetings, as many as you like.
Casual speech, hesitations and “uh-huhs” are all captured accurately. Conversations that mix Japanese and English are transcribed as they are.
Also supports English, Mandarin, Cantonese and Korean. On iOS and Mac, German, French, Spanish, Portuguese and Italian are added, for 10 languages in all.
Each speaker gets their own color. People who have registered their voice are named automatically.
Your recordings are transcribed without going to a server, so it’s safe for confidential meetings. Anything leaves your device only when you choose to share, publish or use AI proofreading.
Field notes
For each conversation, when, where, with whom
and what was said become a single note.
Your notes build up,
and you can look back by person or by date.
Hand your notes straight to an AI agent
to summarize them or plan what’s next.
Field notes / 2026 / September
A meeting with a client. The first delivery date and the inspection plan, recorded along with who said what.
What was said
New in the beta
Every recording appears as a pin where you started it.
iPhone · Mac · AndroidPick a story from a recording and turn it into manga, drawn from portraits of the people in it.
iPhone · Mac · AndroidPublish a set of recordings so people can listen in their podcast app.
iPhoneListen back with large captions while the map shows where you walked and the photos you took.
iPhoneListen back with background noise reduced. Transcription still uses the original audio.
iPhone · MacWhen two people were merged into one speaker, listen line by line and separate them.
iPhone · MacAI proofreading and making cards (manga, slideshows and translations) require signing in and use credits. Transcription stays free, with no time limit.
Community
The Washa community,
sharing how they use it.
Our beta test is running in our Discord community.
Members share what conversations they record
and how they use them, and the team
improves Washa based on what they hear.
Hirano
Interviews on a shopping street
Keeping shopkeepers’ old stories, speaker by speaker
High school student
Grandma’s stories
Keeping family stories in their own voices
Komomo
Reporting at a fishing port
No more note-taking — just listening
Café owner
Chats with regulars
Picking up menu ideas from small talk
Kamii
Dev retreat discussions
Reviewing decisions along with who made them
University seminar
Seminar discussions
Checking later who raised which idea
Gaku
Workshop interview
Turning a craftsman’s words straight into a draft article
With a farmer
Talk of the rice paddies
Reading back what was said during the work
Let us hear your voice, too.
The voices you share make Washa’s listening
more and more precise.
Why not build Washa
together with us?
Enter your email and we’ll send you a link to get Washa right away.