# Using the WebUI — face-to-face meeting recordings

**Any screen not listed here does not exist.** Headings quote the screen's own wording.

**Part of the meetings guide. Other topics:**
conversations, the input box, and sign-in: `references/webui.md`;
documents, files, tables, memory, channels, and apps: `references/webui-workspace.md`;
editing a file in place (「編集」): `references/webui-file-edit.md`;
SIROE pages, blocks, and boards: `references/webui-siroe.md`;
settings, administration, models, and triggers: `references/webui-settings.md`;
external service connections: `references/webui-connections.md`;
`/office`: `references/webui-office.md`;
`/meetings`: `references/webui-meetings.md`;
`/calendar`: `references/webui-calendar.md`.

## 対面の会議の録音(ネットにつながっていなくても録れる)

**For a face-to-face meeting with no video call, including places with no internet connection.** On `/meetings`, 「対面で録音」 (next to 「新しいミーティング」; neither is shown before meetings are set up or in 「ゴミ箱」) opens `/meetings/in-person` (「対面の会議を録音」). Enter a 「タイトル」 (left empty, the date and time you started becomes the title), choose 「公開範囲」 — 「自分のみ」 (the default), 「グループ」, or 「テナント全体」 (it cannot be changed after uploading, and cannot be changed while recording). **「グループ」 appears only after you have opened `/meetings` online on this device while signed in** — the groups you can choose are the ones remembered then (the page itself never goes online); choosing it shows a 「グループ」 field to pick which group. Set 「AI アクセス」 (ON by default; turn it OFF before recording if the meeting must not be transcribed — it cannot be changed while recording, and turning it OFF after uploading is usually too late, because transcription starts right away). **With 「AI アクセス」 ON, choose 「議事録の形式」 (打ち合わせ, 定例会議, ブレスト, 振り返り, 勉強会, 採用面接, 1on1, 問い合わせ対応, 雑談, その他 — the default) and 「会議の言語」 (日本語 by default) under it**: when the device is online and signed in, typing the title fills them in with a suggestion marked 「タイトルから提案」 (a field you chose yourself is not overwritten); with no connection you choose them by hand. Without 「ライブで構造化する」 they cannot be changed while recording (with it ON, 「AI の設定」 changes them; see below), and **a meeting made only when its recording is uploaded (recorded with no connection, or with 「ライブで構造化する」 OFF) can never change them afterwards**, so set them before you start. The meeting language decides the language the uploaded recording is transcribed in and the language of the minutes. Then press 「録音を開始」, and press 「録音を停止」 when the meeting ends. It records **this device's microphone only** and saves to this browser as it goes, so if the tab closes, what was recorded so far is kept; nothing is sent while recording. **After stopping, if the device is online and signed in as the person who recorded it, the page uploads it right away**: it shows the row under 「この端末の録音」 while uploading, then 「アップロードしました。詳細ページで文字起こしと議事録を確認できます。」. Otherwise (no connection, or not signed in) it says 「端末に保存しました。ネットにつながった状態でミーティングの画面を開くと、自動でアップロードします。」. **While recording, the settings are folded away**: the page shows only the title, what was chosen (e.g. 「自分のみ · AI アクセス ON」), a large timer, a ring that moves with the microphone level, and 「録音を停止」. After stopping it shows 「録音を保存しました」 with the message below; 「もう一度録音」 returns to the settings (the title stays). Either way, 「ミーティングの詳細を開く」 appears under the message and opens that meeting's page; if the recording has not been uploaded yet, the meeting page uploads it first (showing the same row) and then shows the meeting. One recording is up to 60 MiB (about 5.8 hours); at the limit it stops and keeps what was recorded. **Do not leave the page while recording** — the link back (「ミーティングの画面へ」) is hidden until you stop. **If the page reloads while recording** (pulled to refresh, F5, or the browser reloading the tab), it comes back to the same recording within 2 minutes: the timer counts from when recording started, the recording after the reload is joined to the one before it into a single recording, and with 「ライブで構造化する」 ON the live panel reconnects to the same meeting. Up to the last 5 seconds before the reload may be lost. If the microphone cannot be used again, it shows 「録音が止まっています」 with 「録音を再開」 (continues the same recording) and 「録音を停止」. After 2 minutes, or after closing the tab, the recording stops and is uploaded as usual. **While recording, the screen does not turn off by the device's auto-lock** (in browsers that support it, including Safari on iPhone with iOS 16.4 or later); turning the screen off with the power button can still stop the recording. Each recording becomes its own meeting.

**Structuring it live while recording (online only):** when the device is online, you are signed in, and 「AI アクセス」 is ON, the page also shows 「ライブで構造化する」 (ON by default; it cannot be changed while recording, and it is not shown with no connection). Recording with it ON creates the meeting as soon as recording starts and shows 「ライブの構造化」 with the tabs 「情報整理」, 「図解」, and 「用語解説」 — topics and decisions, a diagram, and terms explained — to the right of the recorder on a wide screen and below it on a narrow one. **On a narrow screen (a phone) while recording with it ON, the recorder folds into a bar at the top** — the title, the elapsed time, a small microphone that moves with the level, the 「小窓で表示」 icon, and 「停止」 (what was chosen, such as 「自分のみ · AI アクセス ON」, is not shown) — with the captions (the latest few lines) and 「ライブの構造化」 below it, all on one screen without scrolling the page; only 「ライブの構造化」 scrolls inside itself. **The live transcript does not tell speakers apart**; once the recording is uploaded, it is replaced by the transcript with 「話者1」, 「話者2」, … and the minutes are made again from it (if the uploaded recording cannot be transcribed, the live transcript and minutes are kept). Until it is uploaded, the meeting shows 「録音中」 in the list and on its page, and its meeting record shows 「録音中です。録音をアップロードすると、ここで再生できます」 under 「音声」. **A meeting left at 「録音中」 (for example, the recording device was lost) can be deleted by moving it to the trash** — 「ゴミ箱へ移動」 from the row's 「…」 in `/meetings` or from 「その他の操作」 on its page (creator and administrators only); while it is in the trash, its recording cannot be uploaded. If the connection drops, it reconnects by itself; if it is still not back after about 10 seconds, the panel shows 「ライブの構造化に再接続しています。録音は続いています」 until it is. If the live transcription stops (「ライブの文字起こしが止まりました。録音は続いています」), the meeting ends, or the usage limit is reached, it stops with a message such as 「ライブの構造化が止まりました。録音は続いています」 — **the recording itself keeps going**. If the usage limit is already reached, or the live part cannot start, the reason is shown instead of the tabs and it records as usual.

**Sections, fixing, captions, and translation while structuring live:** with 「ライブで構造化する」 ON, the page also shows under it the same section switches as 「新しいミーティング」 — 「情報整理」, 「図解」, and 「用語解説」 (on by default) and 「進行サポート」 (off by default; what to ask next and what to move forward) — and 「情報整理の更新間隔」 (「15 秒ごと」〜「90 秒ごと」, default 60). A section that is off has no tab. **While recording, 「AI の設定」 (the sliders button at the top right of 「ライブの構造化」 on the page; not in 「小窓で表示」) changes the sections, 「情報整理の更新間隔」, 「議事録の形式」, and 「会議の言語」**; the change applies at once, and if it cannot be saved the field goes back and shows why. **Items in 「情報整理」 can be fixed or removed while recording**, as in a call (「修正」 or 「削除」; see 「情報整理」 in `references/webui-meetings-ai.md`), also in 「小窓で表示」. After stopping, neither 「AI の設定」 nor fixing is available. While recording live, a 「字幕」 panel above the tabs shows what this device's microphone hears as captions — words appear as they are spoken, **without speaker names** (one microphone picks up everyone). Before the first words it shows 「話し始めると、ここに字幕が出ます」. 「字幕を隠す」 hides it and 「字幕を出す」 brings it back; the choice is remembered on this device (shown by default). When the meeting's 「翻訳字幕」 is ON (a row in the section switches before recording and in 「AI の設定」 while recording; new meetings start with it OFF, and only the creator and administrators can switch it), choose a language in 「翻訳字幕」 next to it to translate into (英語, 中国語, and so on — the same choices as in video calls): the translation appears under each finished line; the choice is shared with video calls on this device. Translation runs only while the captions are shown, and choosing the meeting's own language shows no translation. The captions and translation also appear in 「小窓で表示」. There are no captions with no connection or with 「ライブで構造化する」 OFF.

**「小窓で表示」 (while recording):** opens a small always-on-top window with the elapsed time, a quiet-microphone warning, and 「録音を停止」 — plus the 「ライブの構造化」 tabs when that is ON — so you can show slides on the same PC and still watch the recording. The page shows 「小窓に表示しています」 meanwhile; press 「小窓を閉じる」 or close the window to bring it back. Stopping from the small window closes it and returns to the page's usual after-recording message. The button appears only in browsers that support it (Chrome and Edge on a computer; not Safari or Firefox).

**Opening it with no connection:** this works only in a browser that has already opened Samurai App online while signed in. Opening any other page with no connection shows 「ネットにつながっていません」 with 「対面の会議を録音する」 (opens the recording page) and 「再読み込み」. Meeting lists, minutes, and past records cannot be read offline.

**Uploading later:** once online, open `/meetings` (opening the meeting's own page also uploads its recording from this device). Recordings still on this device are listed in 「この端末の録音」 above the list and **upload automatically, one at a time** (also when the connection comes back). Each row shows the title, start time, size, the chosen 「公開範囲」, and a state: 「アップロード待ち」, 「アップロード中」, 「オフライン。接続が戻るとアップロードします」, 「別のタブで録音中」, or 「アップロードできませんでした」 with 「再試行」. A 「グループ」 row shows the group's name; if you are no longer in that group when it uploads, it shows 「グループから外れています」 with 「自分のみでアップロード」 (it never widens to 「テナント全体」). An uploaded recording is removed from the device. **If the upload is interrupted, it resumes without making a second meeting.** **A recording made while another account was signed in (or when it is not known who recorded it) is never uploaded automatically:** it shows 「別のアカウントの録音」 with 「このアカウントでアップロード」 and 「端末から削除」 (deleting from the device cannot be undone). **A recording made in a different workspace is not uploaded automatically either** (the same account can belong to several workspaces): it shows 「ワークスペース「<名前>」の録音」 (or 「別のワークスペースの録音」, or 「どのワークスペースの録音か分からない」 for recordings made before the workspace was remembered) with 「このワークスペースにアップロード」 and 「端末から削除」. To put it in the workspace where it was recorded, switch to that workspace and open `/meetings` again.

**The uploaded meeting** appears in the list and opens on `/meetings/<ミーティング ID>` with the 「対面」 badge. **It is visible to the 「公開範囲」 chosen before recording (「自分のみ」 = only the person who uploaded it), and its visibility cannot be changed afterwards**; the recording, minutes, and transcript files have the same visibility. It has no call: no 「参加する」, guest link, recording buttons, participants, 「サム丸招待」, or 「ミーティングを終了する」. Renaming, folders, the trash, 「AI アクセス」, and 「AI チャット」 work as for other meetings. The meeting record screen shows 「音声」 with a player in place of the video recording, plus the transcription state: 「文字起こしを待っています」, 「文字起こし中」, 「AI アクセスが OFF のため文字起こししていません」, 「今月の利用上限に達しているため、文字起こししていません」 (the creator's usage limit was reached when it was uploaded; the recording is kept), or why it failed (for example a silent recording). **When transcription finishes, the transcript and minutes appear as for a call** (see 「議事録と文字起こし」 in `references/webui-meetings.md`; a long recording takes a few minutes, and the page keeps checking until the minutes arrive, even when the recording was uploaded hours after it was made). Speakers are told apart by voice and named 「話者1」, 「話者2」, …. **The meeting's creator or an admin can name them**: in 「話者の名前」 at the top of the transcript, press a speaker, type a name (up to 40 characters), and press Enter or click elsewhere; save it empty to go back to 「話者1」. The name replaces the number in the transcript and its search, the minutes, the transcript file, and 「AI チャット」 — the minutes and transcript files are rewritten (minutes a person edited are left as they were, and new minutes are made). Others see the names but cannot change them. With 「AI アクセス」 OFF before transcription, only the recording is kept (no transcript and no minutes); if it is turned OFF after transcription, a name changes only the screen and the minutes and transcript files are not rewritten.

**「ミーティングメモの拡張」 (the browser extension that records Google Meet / Zoom / Microsoft Teams):** a Chrome / Edge extension that records a Google Meet, Zoom, or Microsoft Teams meeting in the browser (tab audio plus the microphone), transcribes it, and saves it to this workspace's ミーティング. **Install and connect it from the 「ミーティングメモの拡張」 button at the bottom of the meetings list's sidebar** (below 「固有名詞の辞書」 and 「議事録を通知先へ送る」; not shown while meetings are not configured for the workspace). The dialog shows 「ストアを開く」, the extension ID, and, for admins, the line to add to the Chrome policy `ExtensionInstallForcelist` for managed devices. **It cannot record until connected** — press 「ミーティングメモの拡張をつなぐ」 (it can be pressed only once the extension is installed) and finish signing in in the window it opens; recording then starts from the extension's side panel. **Before pressing 「記録を始める」, the side panel lets you choose the meeting's settings**: 「公開範囲」 (「自分のみ」, 「ワークスペース全体」, or 「グループ」 with the group to share with — it cannot be changed after the upload), 「AI アクセス」 (OFF stops transcription, minutes, and AI), 「議事録の形式」 (「打ち合わせ」 by default), 「ミーティングの言語」, 「ミーティングの分析」 (OFF by default), and 「ライブで構造化する」 with the panels (「情報整理」, 「図解」, 「用語解説」, 「進行サポート」), 「翻訳字幕」, and 「情報整理の更新間隔」. The extension remembers the choice for the next recording. **「ミーティングの名前」 is filled in from what the meeting tab shows** — the Google Calendar event's title in Google Meet, the meeting's name in Microsoft Teams — and can be edited before starting. Zoom, and Google Meet meetings without a calendar event, leave it empty; if it is empty when recording starts, the meeting is named after the service and the start time (e.g. 「Google Meet 2026-10-26 14:00」). **While recording with 「ライブで構造化する」 ON, the side panel shows the live transcript and one tab per panel that is ON** — the same 情報整理, 図解, 用語解説, and 進行サポート as in face-to-face recording on the web: 「AI の設定」 above the transcript turns panels and 「翻訳字幕」 on or off and changes 「情報整理の更新間隔」, 「議事録の形式」, and 「ミーティングの言語」 while recording; 「字幕を隠す」 / 「字幕を出す」 hides the transcript, and with 翻訳字幕 ON the select next to it picks the language to translate into on this device; each panel has a copy button; items in 情報整理 can be fixed or deleted with the pencil and trash buttons; 図解 can be zoomed and exported as HTML. The red dot next to 「記録中」 grows with the microphone volume, and a notice appears while your voice is too quiet to pick up. **While recording, 「スクリーンショットを撮る」 in the side panel saves what the meeting tab shows** to the meeting's 「スクリーンショット」 tab; the meeting tab must be the one shown in its window (otherwise it answers 「記録しているミーティングのタブを表示してから撮ってください」), and if it asks to reopen the side panel, press the extension icon on the meeting tab and try again. 「つなぎ直しが必要です」 means the sign-in expired; 「別のワークスペースにつないでいます」 means meetings are saved to that other workspace — pressing 「ミーティングメモの拡張をつなぐ」 switches it to this one. Other browsers (Firefox, Safari) cannot use it.

**Meetings recorded by the browser extension (Google Meet / Zoom / Microsoft Teams):** a meeting held in Google Meet, Zoom, or Microsoft Teams in the browser and recorded with the browser extension is saved as a face-to-face meeting, but the list and its page show the service name — 「Google Meet」, 「Zoom」, or 「Microsoft Teams」 — instead of the 「対面」 badge. While it is being recorded, it shows 「録音中」, and its meeting record shows 「録音中です。拡張で記録を停止すると、ここで再生できます」 under 「音声」: the extension uploads the recording when you stop recording, so there is nothing to upload by hand. Unlike other face-to-face meetings, it has a 「スクリーンショット」 tab (screenshots taken from the extension's side panel). The live transcript uses the service name as the speaker; after the upload it is replaced by 「話者1」, 「話者2」, … like any face-to-face recording.