Sign in

Airffic Minutes

Voice to text

Turn speech into text as you talk, and let AI pull out decisions and next actions.

The app below is currently in Japanese. An English interface is on the roadmap; everything you enter stays on your device until you sign in.

音声はこの端末の中で文字にします。外には文字だけを送ります。認識の言語：日本語（「設定」で変えられます）

0 文字

要点

例決まったこと・次にやること・持ち越し、の順で出ます

入力：30 分の定例（文字起こし 2,400 文字）

決まったこと ・10/10 の外壁点検は午前 9 時集合。機体は 2 台 ・報告書は前回の形式。写真は南面から順に 次にやること ・山田：見積の再提出（9/25） ・鈴木：居住者への掲示文を作る（9/27） ・佐藤：屋上の鍵の手配を管理会社に確認 持ち越し ・冬季の点検日程は次回に

出力の形を示す例です。実際の答えは、書いた内容で変わります。

What Airffic Minutes is

What you want after a meeting is not the recording but what was decided and what happens next. Airffic Minutes turns speech into text as you talk, and AI organizes the key points as decisions, next actions and carried-over items. Pasting an existing transcript works the same way.

Speech recognition is a browser feature (Chrome and Safari) and runs on your device. Only text leaves the device; audio is never sent. Key points are produced by Anthropic's Claude. Saved minutes can be searched by title, key points or full text, and every "next action" is gathered into one to-do list.

There is no loading of long recordings, no speaker separation and no integration with video-call services. It is for the meeting in front of you, turned into text and key points on the spot. The summary is currently written in Japanese. For summarizing long text alone, use Airffic Summary. The app itself is currently in Japanese; an English interface is on the roadmap.

How to use it

  1. 1RecordPress "record" and speak. Text appears as you go. Or paste a transcript.
  2. 2Extract key pointsGive it a title and press "key points". Decisions, next actions and carry-overs appear.
  3. 3Save and search"Save" adds it to your minutes. Search by title, key points or full text.
  4. 4Work through the to-dosEach meeting's next actions gather under "to do". Tap to clear them when done.

Questions

Is audio stored or sent?
No. Audio becomes text inside the browser, and only text leaves the device. What is saved is also text: the transcript and the key points.
Which browsers are supported?
Speech recognition works in Chrome and Safari. In other browsers you can still paste text and extract key points.
Can it tell speakers apart?
Not yet; it produces one continuous text. Speaker separation is a future candidate.
Can I use it for meetings in English?
Yes; switch the recognition language to English in settings. The key points are currently summarized in Japanese.
Do I need to sign in?
Transcription works without signing in. AI key points and saving to your account require sign-in (without it, saves stay on this device).

Related