kensaurus

Field note · Speech to timed text

Keep a conversation as timed text, with no account: Hear, Hear

Hear, Hear turns speech into timed text in your browser: reread it, export TXT, Markdown, DOCX or SRT. Takes are saved only in this browser. Free, no sign-up.

Kenji Sakuramoto · ｋｅｎｓａｕｒｕｓ合同会社 ·

"So the launch is Friday — that's confirmed." On the Hear, Hear home page, a sample meeting types itself out line by line, each line stamped with the moment it was said.

Hear, Hear is a browser app that turns speech into timed text you can reread, note tasks from, and export. It is there so a conversation can be kept without a paid transcription account. There is no account at all: you open the page and record.

The Hear, Hear home page in Japanese: the wordmark HEAR, HEAR. in huge letters over ruled lines, and a sample meeting transcribing itself at the top right, two speakers at 00:03 and 00:07 confirming a Friday launch. At the bottom left, a Speak button with a note that the voice goes to the browser's speech recognition (Google, in Chrome) and nothing is saved there
The home page. The meeting is sample data (demo, captured 2026-10-02)

Record, reread, take it with you

In the Record room you press the red bar and talk. Text appears as you speak, in order. You can record from the mic, and in desktop Chrome and Edge also from screen audio, or both at once. In this public version a take can run up to five minutes.

The Record room in Japanese: a poster-size 00:00 timecode with a five-minute demo limit, and a full-width red Start recording bar. On the right: source (mic, screen, both), scene, a microphone permission button, and a note that the demo uses the browser's speech recognition (through Google's speech service in Chrome) and keeps takes and transcripts only in this browser
The Record room. The red bar is the only colour on the page (demo, captured 2026-10-02)

When you stop, the take opens in the reader. Every line keeps its timestamp, and tapping a line plays that moment again. You can export a take as TXT, Markdown, DOCX or SRT, and the file is built in your browser. Decisions and to-dos go in the Tasks room, which you fill in yourself. The whole app switches between Japanese and English.

The reader in Japanese: a take titled with a Friday release check, 20 seconds long, a play bar, export buttons for .TXT, .MD, .DOCX and .SRT, and four timed lines from 00:03 to 00:16
The reader, with export (demo, captured 2026-10-02, sample take)

All three were captured from the demo on 2026-10-02 for the kensaur.us exhibit, with the interface in Japanese. I checked on 2026-10-16 that the live pages still read the same. The meeting and the take are sample data, not real people.

Where your voice goes

Recordings and transcripts are saved only in this browser. The app does not upload them. Live transcription, though, uses the browser's own speech recognition, and in Chrome that means your voice is sent to Google's speech service. Clearing the site's data deletes your takes. If your browser has no speech recognition, the page tells you so.

It's free

All of it is free, with nothing to sign up for and nothing to pay. This public version has no summaries, no AI-drafted tasks, no PDF export and no sync between devices.

See it in the exhibit hall