Vū Intelligence

Vū Intelligence brings live session intelligence into Vū Studio. It follows along with your sessions, captures what happens, and surfaces insights — so you can focus on the conversation instead of taking notes.

Vū Intelligence works from a session's transcript. It can capture that transcript two ways:

• In the room — from the microphone of the space you are in.
• In a virtual meeting — by sending a meeting agent (a bot) to join a Zoom, Google Meet, or Microsoft Teams call.

Either way, the transcript feeds the same pipeline: as people talk, Vū Intelligence builds live insights and a concept graph (a mind map) that you can review during and after the session, and chat with through Vū AI afterward.

Your sessions live at one.vu.studio/ai/sessions. To get there inside Vū Studio, click the Vū Studio icon to go Home, click Vū AI to open the Vū AI module, then click Vū Intelligence — or go directly to one.vu.studio/ai/sessions.

Starting a Session

Start a Vū Intelligence session from Vū AI: open one.vu.studio/ai/sessions (Vū Studio icon → Vū AI → Vū Intelligence), click New session, and give it a Name and an optional Focus. During the session, Vū Intelligence builds a live record you can refer back to afterward.

Creating a Session: Presets, Tags, and Custom Settings

New session at one.vu.studio/ai/sessions opens the New session form. It decides what the session is; how it captures audio comes next.

1. Type a Name (required).
2. If you or your team have saved presets, pick one from Preset. The default preset is already selected and marked — default. Choose No preset — built-in settings to start from the built-in settings instead. The Preset field only appears once at least one preset exists. See Session Presets.
3. Add an optional Focus — what the session should pay attention to, for example budget decisions and customer objections. Picking a preset fills this in; you can type over it.
4. Add optional Tags to describe the meeting — see Tagging Sessions.
5. Check the one-line summary under the tags. It shows what the session will start with, for example Names anonymized · Images: …. To change something for this session only, open Customize for this session. It holds When this session ends, Anonymize names in insights, Illustrate insights, and Delete the transcript when this session ends, all filled in from the preset.
6. Select Create session.

If you change anything under Customize for this session, a Custom settings badge appears. Reset to preset puts the fields back to the preset's values. Changes made here apply to this session only; the preset itself stays the same.

The session then opens on a short capture setup, Set up "your session name", headed How do you want to capture?:

• This room's mic — record in person from this device. Choose where you are recording, and tick Automatically start recording if the mic should open as soon as the session starts. That choice is remembered on this device. Then select Start Session.
• A meeting agent — paste the meeting link, optionally choose a Present a view (screenshare) option and End the session when the meeting ends, then select Send agent.

Spoken language is at the bottom of the setup; see Choosing the Spoken Language. Outputs, Security, and Focus opens the session's full Session settings and brings you back to the setup when you close them.

Setting Up and Recording in One Step

A brand-new session opens a short setup — Set up "<your session name>" — that asks one thing at a time and ends by starting the recording, so getting set up and getting going are a single pass rather than a configuration form followed by a hunt for the record button. Every step is skippable; sensible defaults apply.

1. What's the focus of this session? — steers what Vū pays attention to as the conversation unfolds. Optional; leave it broad if you like.
2. What should happen when it ends? — one free-text instruction Vū runs for you automatically when the session ends, for example "Write a whitepaper of the session and email it to the team." Shortcut buttons for Whitepaper, Landing page, and Email summary fill it in.
3. Privacy — Anonymize names in insights (refers to people as "Speaker A/B/C" and scrubs names before insights are stored) and Delete transcript when the session ends (the raw transcript is purged after any automation runs; insights and the mind map are kept).
4. How do you want to capture? — This room's mic (record in person from this device) or A meeting agent (send a bot into a Zoom, Google Meet, or Teams call).

On the microphone path the final button is Start recording: it saves the setup and begins capturing in one click. Skip setup takes the defaults instead. Everything here can be changed later from Settings in the session's title bar.

The Record Bar

Inside a session, recording lives in a single pill at the bottom of the screen — the same control creators and joined contributors both see, and it stays put as you move between the session's tabs.

• Record starts capture and becomes Stop. While recording, the words being transcribed stream past in the bar so you can see it is hearing you.
• The gear icon (Audio setup) opens the audio panel. It asks one plain question — Where are you recording? — and maps your answer to the right capture settings:

Choice	Best for
In-person room	A meeting or workshop in one space — picks up voices across the table and labels who's speaking.
Online meeting	A video call — captures the meeting tab's audio plus your mic. Chrome or Edge only; use headphones.
Just me	You speaking into a headset or mic — fast and clean, no speaker labels needed.

• Fine-tune audio settings opens the underlying microphone controls (input system, plus the raw echo-cancellation, noise-suppression, and gain toggles) if a room needs hand-tuning.
• Session creators also get Add a meeting agent in the same panel, for capturing an online call instead of the room.

Changes made while a recording is running apply the next time you start recording — press Stop, then Record again, for a new choice to take effect.

The Audio Setup Panel

In the current version of Vū Studio, the Audio setup (gear) panel in the record bar, and the This room's mic option when you set up a new session, show the same short set of controls:

• Also capture an online meeting — adds the audio of a call running in another browser tab to your mic. You pick the tab when recording starts; see Capturing an Online Meeting. This option is not shown in browsers that cannot share tab audio, such as on a phone or tablet.
• Fine-tune audio settings — opens Mic environment, with Room (Processing off + gain up — picks up quieter/distant voices across a room.) and Headset (Noise suppression on — best when speaking straight into the mic.), plus individual Noise suppression, Echo cancellation, and Auto gain switches.
• Add a meeting agent (session creators only, in the record bar) — Sends a bot into a Zoom, Meet, or Teams call. No mic needed. See Bringing a Meeting Agent Into a Virtual Call.

Capturing an Online Meeting

Online meeting captures the meeting tab's audio alongside your mic, so both halves of a video call land in the same transcript. How browsers handle tab audio shapes the steps:

1. Have the call open in its own browser tab, then press Record in the session.
2. Your browser opens its sharing picker. Choose the tab the call is running in, and switch on the picker's tab-audio option (Chrome labels it Also share tab audio) before you confirm — without it the tab is shared silently and none of the call is captured.
3. Your browser then shows its own sharing this tab bar for as long as the capture runs. Leave it in place; ending the share ends the meeting half of the capture.

If the share does end — you press the browser's Stop sharing, or close the tab — recording stops and the record bar tells you: Meeting capture stopped — sharing ended. Press Record to pick the meeting's tab again. Re-opening the picker requires a click, so the session deliberately does not try to restart this capture on its own.

If the browser cannot supply diarized meeting audio, the session records this room's mic rather than failing outright, and says so: Recording this room's mic only — the online meeting's audio couldn't be included. Use Chrome or Edge on a desktop for meeting capture.

To capture a call without giving up a browser tab — or from a phone or tablet, which cannot share tab audio at all — send a meeting agent instead; see Bringing a Meeting Agent Into a Virtual Call.

Live Transcription and Speaker Labels

Speech is transcribed as you talk — a live interim ticker while you speak, then a finalized line a moment later — and finalized lines carry speaker labels where the chosen input system supports them, so the transcript reads as turns (Speaker A, Speaker B, …) instead of one undifferentiated stream. Those labels also sharpen the insights and the mind map built from the transcript.

• In-person room and Online meeting produce speaker labels.
• Just me uses the browser's own fast recognizer and produces no speaker labels.
• Labels are letters, never names — the same anonymization the Anonymize names in insights toggle applies. (Names captured from a virtual meeting are the exception; see Automatic Sessions from Your Calendar.)

Separate Audio Sources

When one recording has more than one source of audio, Vū Intelligence transcribes each source separately instead of mixing them together, so two people talking on different sources are not merged into one speaker. This happens automatically; there is no setting.

What you are recording	How speakers are labelled
One microphone	Speaker A, Speaker B, …
Your microphone plus an online meeting (Also capture an online meeting)	Room · A, Room · B, … for people picked up by your microphone, and Call · A, Call · B, … for people in the call
An audio interface or microphone array with three or more input channels	Mic 1 · A, Mic 2 · A, … — one number per input channel

Letters restart for each source, because the source name already tells them apart. A two-channel (stereo) microphone is treated as one microphone. When a single session runs long enough to use every letter, labels continue as AA, AB, and so on, so a label is never reused for a different person.

If the transcription service cannot accept the sources separately, recording carries on with them mixed together, labelled as for one microphone.

Choosing the Spoken Language

Each session is transcribed in one Spoken language. It is English unless you choose otherwise, and the same choice shapes the insights, the report, and the dashboard, which are written in that language too.

You can set it in three places:

• in the capture setup that opens when you create a session (Spoken language, at the bottom);
• in the session's Session settings (the settings button in the session's title bar), while the session has not ended;
• in a preset, so every session started from it uses that language — see Session Presets.

A change applies the next time recording starts, and to every device that joins the session.

Multilingual (mixed languages), at the end of the list, is for rooms where people really do switch language mid-conversation. Use it only then: it can mistake an unclear moment for another language entirely. If a chosen language cannot be used, transcription carries on in English rather than stopping.

Names and Jargon

Transcription most often gets names, product names, customers, and acronyms wrong. To help it, open Session settings and fill in Names and jargon with the words to listen for, separated by commas — for example Art-Net, SEND, Northwind. The field is available until the session ends, and the words listed under Listening for: are the ones in use.

If you leave it empty, Vū Intelligence picks names and acronyms out of the session's Focus instead. Keep the list short and specific: a word that is boosted is more likely to be heard everywhere, including where it was not said.

Chat Counts as Session Input

Anything typed during a running session is treated as part of the session, not as a side conversation. That covers both places people type:

• the chat inside the online meeting itself, in Google Meet, Zoom or Teams, where a meeting agent has joined the call;
• the session's own Chat drawer in Vū Studio.

Typed messages are written into the transcript alongside the spoken lines, attributed to the person who sent them and marked as chat rather than speech. Because insights are built from the transcript, chat reaches the insights automatically — so a link someone drops in the chat, a name spelled out for the record, or a correction typed while someone else is talking all count.

Vū Intelligence is told that these lines were typed rather than said, and reads them accordingly: as background context handed over deliberately, and as a signal in their own right. A question asked in the chat often means something spoken was unclear, which is frequently worth an insight by itself.

Two things worth knowing:

• Only what people type is captured. Vū Intelligence's own replies in the chat are never written into the transcript.
• Typing is not microphone time. Chat lines do not count toward the session's recorded minutes, do not keep a finished session from closing out, and do not put a phantom microphone on the device list. Only speech does.

Privacy works exactly as it does for speech: chat lines are transcript lines, so Anonymize names in insights and deleting the transcript at the end of a session both cover them with no separate setting.

note

The suggestion chips in the session chat drawer (for example Summarize the session so far.) are chat messages like any other, so pressing one during a live session writes that sentence into the transcript as a typed line.

Recording on a Phone or Tablet

Recording works from a mobile browser, iPhone and iPad included: open or join the session on the phone, allow microphone access when prompted, and use the same Record control. Two things to know:

• The Online meeting option captures another tab's audio, which mobile browsers cannot do, so it is hidden there.
• A microphone never picks up the remote participants of a call you are listening to — their audio comes out of the speaker rather than into the mic. To capture an online meeting, send a meeting agent instead.

Bringing a Meeting Agent Into a Virtual Call

1. Open the session at one.vu.studio/ai/sessions (Vū AI → Vū Intelligence) and go to the Devices tab.
2. Select Add meeting agent (it sits alongside Add devices).
3. Paste the meeting link for your Zoom, Google Meet, or Teams call.
4. Optionally choose a view to Present a view (screenshare) and whether to End the session when the meeting ends, then select Join meeting.
5. Vū Intelligence dispatches a bot that joins the meeting and begins capturing the transcript. The bot appears as a device row on the session, labelled by platform and marked in call, so you can remove it at any time (removing it makes the bot leave the call).

As the meeting proceeds, insights and the mind map update live — you do not need to wait until the call ends.

Presenting a View Back Into the Call

You can have the meeting agent present the session back into the meeting itself. This is opt-in: in the Add meeting agent dialog, the Present a view (screenshare) control is set to Off by default, and you can switch it to Overview, Mind map, or Flow. Anyone in the call can also change it later by @mentioning the agent in the meeting chat.

The screen share and the agent's camera tile are deliberately different, so the two feeds are not redundant: the screen share carries the view you picked, while the camera tile shows a compact branded Vū Intelligence view.

Talking to Vū Intelligence during the call

Participants can @mention Vū Intelligence in the meeting chat to ask it questions about the session. It replies in the chat, grounded in what the session has captured so far.

• On Microsoft Teams, use the chat's native @mention — start typing @ and pick Vū Intelligence from the suggestions. The mention is recognized even when it lands in the middle of a message.
• On Google Meet (which has no mention autocomplete), type the mention as plain text — @vu, /vu, or @Vū Intelligence — at the start of your message.

Ending the session

End the session when the meeting is over to finalize its insights. There is also an optional auto-end setting that ends the session automatically when the meeting ends (off by default).

Capturing names and emails

Vū Intelligence always captures participant names, and captures emails when they are available. On a plain link-join, participant emails are often not available — connecting your calendar (see below) is what reliably supplies attendee emails for follow-up automations.

New Session Defaults

Every new Vū Intelligence session starts from a set of defaults. You can change them, and if you own or administer a team you can set them for everyone in that team.

Where they live: https://one.vu.studio/ai — click the Vū Studio icon to go Home → click Vū AI to open the Vū AI module → open Settings → choose the Intelligence Sessions tab.

Changing a default never affects a session that already exists. Each session keeps the settings it was created with.

Whose defaults apply

• A team owner or admin editing this tab is editing the team's defaults. They apply to every session anyone creates in that team and override each member's own defaults there. The form tells you which team you are editing.
• A team member editing this tab is editing their own defaults, used for their personal sessions. Sessions they create in the team use the team's defaults instead, which only an owner or admin can change.

The settings

Setting	What it does
Focus	What sessions should pay attention to — for example, budget decisions and customer objections. Leave it blank to set the focus per session instead.
When a session ends	An instruction Vū AI runs automatically once a session finishes — for example, write a whitepaper of the session and email it to the team. Example prompts are offered as one-click starting points.
Anonymize names in insights	Refers to people as Speaker A, Speaker B, Speaker C and scrubs names before insights are stored.
Illustrate insights	Adds an image to the occasional insight worth illustrating. Choose how often from the dropdown; each option explains itself underneath.
Delete the transcript when a session ends	Purges the raw transcript when the session ends, after any end automation has run. Insights and the mind map are kept.
Spoken language	The language sessions are transcribed in, and that insights, the report and the dashboard are written in. See Choosing the Spoken Language.

Click Save to apply. The button names the scope you are saving, so you can see whether you are writing your own defaults or the team's.

Session Presets

A preset is a named, reusable setup for a new session — useful for a meeting that runs the same way every week. The Intelligence Sessions tab of Vū AI Settings (one.vu.studio/ai → Settings → Intelligence Sessions) lists them under Session presets. The preset marked Default is used whenever you create a session without picking another.

To create a preset:

1. Select New preset.
2. Give it a Name, for example Weekly standup.
3. Fill in the settings: Focus, When a session ends (the Try: buttons — Whitepaper, Email summary, Landing page — fill in a starting prompt), Anonymize names in insights, Illustrate insights, Spoken language, and Delete the transcript when a session ends.
4. Select Create preset. Cancel discards it.

To manage presets:

• Edit opens a preset; Save preset keeps the changes. A preset you are not allowed to change opens with View instead, read-only, with Back to return.
• The delete button asks Delete "preset name"? — confirm with Delete.
• Owners and admins set the default with the radio button beside a preset.

Adding, deleting, and changing the default take effect at once. A scope can hold up to 20 presets.

Whose presets you see. The tab shows one scope at a time: the active team's presets when you are working in a team, and your own otherwise. Team sessions use the team's presets and personal sessions use yours; the two are not mixed.

• Anyone in a team can add presets to the team's list, and edit or delete the ones they created.
• Only a team owner or admin can edit other people's presets or change which preset is the team's default. If no default is set, new sessions start from the built-in settings until someone picks a preset or an owner or admin sets one.

Presets apply only when a session is created. The values are copied onto the new session, so editing or deleting a preset later never changes a session that already exists. To start one session differently, pick another preset — or No preset — built-in settings — in the New session form, or use Customize for this session there. See Creating a Session: Presets, Tags, and Custom Settings.

Devices and Breakout Rooms

Everything and everyone contributing audio to a session — phones, laptops, walls, and meeting agents — is managed from one place: the Devices panel on the session page. On the session page it is one of the tabs across the top; elsewhere it appears as an expandable Contributors and Viewers section you click to open, with a count of the contributors on it.

Each device row shows a presence dot (recording now, idle, or stopped), the device's name, and how much it has captured. From the row you can:

• Rename it — click the name and type. Useful when four identical "Chrome on Mac" entries turn into Front table, Back table, Wall, and Priya's phone.
• Stop or allow its recording — the stop button halts that device's capture remotely. It is not a silent cut: the device itself shows The host stopped your recording and stays locked out until you allow it again, so nobody is left wondering why their microphone went quiet. The play button lets it record again.
• Remove a meeting agent — the remove button on an agent row makes the bot leave the call, and asks you to confirm with Leave first. Microphone devices have no remove action; stop and allow are the whole model for them, so a device you stop can always be brought back.

Your own audio appears as a Host row, marked This device.

Adding Devices

Add devices opens a link you can share with anyone who should join:

• Choose Collaborate (view plus add their mic) or View only (watch, no mic).
• Choose which view the link Opens in — Full view, Mind map, or Flow.
  • Full view is the normal live view: the mind map, the insights sidebar, and the record bar. It is the default for a Collaborate link, so someone you invite to add their microphone lands on a page they can actually work in.
  • Mind map and Flow are the chrome-less presentation layouts, intended for a wall or a cast. Mind map is the default for a View only link.
  • Cards is the insight board with the sidebar, so the transcript stays visible alongside it. See The Cards View.
• Share it by Copy, by scanning the QR code from a phone, or by typing an address into Email the link to… and selecting Send.

If the session belongs to a team, Only my team can open shared links requires anyone opening these links to be signed in and a member of that team. With it off, anyone holding the link can open it.

Breakout Rooms

Turn on Breakout rooms at the top of the panel when you want contributors split into focused groups instead of one merged conversation. The tab label changes from Devices to Breakouts, and the panel changes with it:

1. Select Add room to create a room, then click its name to rename it.
2. Give the room a focus in the text box beneath its name — for example "Explore pricing objections." Each room's focus steers the insights generated from that room's audio, layered on top of the session's overall focus.
3. Select Get link on the room to share a join link scoped to that room: a device that joins through it is placed in the room automatically.
4. Move a device between rooms with the dropdown on its row; devices that have not been placed yet sit under Unassigned.
5. Agent on a room row sends a meeting agent into that room's call. One agent per room.

Insights and the mind map stay tagged per device, so you can still filter by group. Deleting a room asks you to confirm with Delete.

A room's focus is set by the person who owns the session, from this panel. Someone joining through a room link does not get a focus field of their own — one room has one focus, steered from one place.

A link created with Get link on a room row is scoped to that room. A device that joins through it is placed in that room and stays there; it cannot pick a different room for itself.

Filtering Insights by Breakout Room

When breakout rooms are on and there is more than one room to choose between, a Filter by breakout room dropdown appears above the insights list — in the sidebar on a desktop, and at the top of the Insights tab on a phone.

1. Open the session at one.vu.studio/ai/sessions (Vū Studio icon → Vū AI → Vū Intelligence) and select the session.
2. Open the Insights view.
3. Select the dropdown at the top of the list and pick a room. Pick the all-rooms entry to go back to the whole session.

The filter narrows the insights to the microphones sitting in that room, so a three-track event can be read one track at a time instead of as one blended conversation. Room membership is read from the live device roster, so moving a device between rooms is reflected the next time the list refreshes. If a room has produced nothing yet, the list says No insights for this breakout room yet rather than appearing empty.

Two Devices in One Room

When two devices capture the same room, both can hear the same person. Vū Intelligence keeps one copy of a sentence that arrives from two different devices within a few seconds of each other, so the transcript, the insights, and the mind map are not built from doubled text.

Only longer lines (five words or more) that are nearly identical are treated as echoes. Short replies such as yes or makes sense are always kept, because different people really do say them at the same time. The first copy to arrive is the one kept.

If Another Tab on This Device Is Already Recording

Vū Intelligence allows one live microphone per device. If you open a join link in a browser where another Vū tab is already recording into a session, that link shows a card headed Another tab here is recording instead of the setup form, and the record control is disabled with the hint Another tab on this device is recording.

This exists because two microphones in the same room capture the same voices twice: the transcript is duplicated, the insights are generated from doubled text, and — because sessions bill per microphone-minute — the room is charged twice for one conversation.

Only you can tell whether the other tab is genuinely the same microphone in the same room, so the rule is advice rather than a lock. Select Record here anyway on that card to override it and start recording. The override lasts for the rest of your visit.

Things worth knowing:

• The check asks whether another tab is recording, not whether one is merely open. Having a session page open without recording does not block anything.
• It only sees tabs in the same browser profile on the same device. A phone, a second computer, or a private/incognito window is never blocked by it.
• It is checked twice — when the page opens and again at the moment you take the microphone — so two tabs cannot both slip through and start together.

When an Invite Link Doesn't Open

Session and breakout-room invite links last 24 hours from the moment they are created. Using a link does not extend it; create a fresh one from the Devices panel when an old one runs out.

If a link does not open, the page now says which of these happened rather than giving one generic message, because the fixes are different:

What the page says	What it means	What to do
This invite link has expired	The link is older than its 24-hour lifetime.	Ask the session owner for a new link.
The session was deleted — there's nothing left to join	The session the link pointed at no longer exists.	Ask the owner whether it moved.
The link is not valid	The link is incomplete or was altered — most often truncated by a chat or email client.	Copy the whole link again, or ask for a fresh one. A QR code avoids the problem entirely.
Something went wrong on our end — the link itself looks fine	A temporary fault checking the link. The link is not the problem.	Select Try again. This is the only one of the four that offers a retry.

If a whole batch of links stops working at once and they are not all a day old, that is worth raising with support rather than reissuing them one at a time — see Get Support.

Casting a Session

Casting puts the live session on a wall, separately from sharing join links. Select Cast in the session's title bar to open the Cast to a wall dialog.

1. Choose what the wall gets: View only (the wall watches the live view) or Contribute (the wall can also add audio from the room it is in).
2. For View only, choose which view it Opens in — Mind map or Flow.
3. Tick one or more of your connected walls in the list, then select Cast.

Opens in also offers Cards, the insight board — see The Cards View. A session that has ended, or a shared session, can still be cast; it casts view-only.

A cast wall shows the immersive view only — no tabs, no toolbar, and no filter or zoom controls, since there is nothing to click on a wall. It also keeps the session's insights generating on its own, so analysis continues even with your own tab closed.

If the list comes up empty, no walls are currently connected — connect to one first, then reopen the dialog.

Automatic Sessions from Your Calendar

Vū Intelligence can automatically create sessions from your calendar events. Once your calendar is connected, upcoming meetings can have a Vū Intelligence session set up for them in advance, so everything is ready when the meeting begins.

This is an opt-in feature. When it is enabled and you have a virtual meeting on your connected calendar, Vū automatically creates an Intelligence Session and dispatches a meeting agent to join the call — with no manual "add meeting agent" step — and captures the event's invitee emails onto the session for use in follow-up automations.

To use it:

1. Open your sessions list at one.vu.studio/ai/sessions (Vū AI → Vū Intelligence). The Auto-join calendar meetings bar sits just under the page heading.
2. Connect a calendar. Google Calendar and Outlook are supported. If neither is connected the bar reads Connect Google Calendar or Outlook to enable and offers a Connect button, which opens the connections screen.
3. Turn on automatic sessions with the switch at the right of the bar (opt-in). Once on, the bar names the connected calendars and confirms that your scheduled video meetings are captured automatically.
4. When an upcoming virtual meeting is detected, Vū creates a session for it ahead of time and sends the agent to join when the meeting starts.

On the sessions list, an auto-created session that has not run yet shows a Scheduled state, and a session that involves multiple people from your organization shows a Shared badge.

Attendee emails come from the calendar invite

The emails captured this way are the meeting's calendar invitees. This is ideal for "email everyone who was invited" style follow-ups. (The bot on its own still cannot read the emails of people who join only by link — the calendar event is the email source.)

What the Meeting Agent Does in the Call

Whether it was sent by hand or dispatched automatically from your calendar, the agent behaves the same way once it is in a Zoom, Google Meet, or Teams call. It creates the session, or attaches itself to the one already waiting for that meeting, and then:

• Ingests the transcript in real time. Insights, the mind map, and the timeline all build during the call rather than after it — open the session in another tab and watch them arrive.
• Carries real participant names through as speakers. Because the meeting platform supplies who is talking, transcript lines and insights are attributed to people by name instead of Speaker A/B. The session's Anonymize names in insights setting still governs what is stored in the insights themselves.
• Captures the attendee roster. Expand attendees captured on the agent's device row to see exactly who it recorded — names always, and emails where the platform or the calendar invite supplied them.
• Can show the session back to the room. Set Present a view (screenshare) to Mind map and the live mind map is screen-shared into the meeting; Flow shares the insights timeline instead.

The Agent's Camera Tile

The agent's camera tile is deliberately simple rather than a second copy of the mind map: a compact Vū Intelligence header, the concept currently being tracked, and the latest insight in a colour-coded card — sized to stay readable in a small meeting thumbnail. Newly joining participants are greeted by name.

The fuller Overview layout is available as one of the optional screen-share modes, alongside Mind map and Flow, for when you want that detail on the shared screen instead of the camera tile.

Reviewing Insights

After a session, review the insights Vū Intelligence captured. Use them to follow up, share highlights, or feed the results into other Vū Studio workflows.

Each session keeps:

• Insights — the key points surfaced from the conversation.
• A concept graph (mind map) — the topics and how they connect.
• The transcript — the underlying record the insights are built from.

You can find your sessions at one.vu.studio/ai/sessions (Vū Studio icon → Vū AI → Vū Intelligence), where you can search by title or focus and filter by state (for example, not started, in progress, or ended).

Finding Sessions by Date (Calendar View)

When the list gets long, browse your sessions by date instead. At the top of one.vu.studio/ai/sessions (Vū Studio icon → Vū AI → Vū Intelligence), switch the view from List to Calendar.

The calendar shows one month at a time:

• Each day lists up to three sessions with their start time and title, coloured by state. A day with more shows +N more.
• Previous month and Next month (the arrows) move between months; Today jumps back to the current month.
• Select a session to open it. Select a day's number — or +N more — to open that day.

The day view lists every session on that date, newest first, with its state and when it Started and Ended. Select one to open it, or Back to month to return to the grid. A day or month with nothing on it says No sessions on this day. or No sessions in the month.

Sessions are filed under the date they started. A session that was created but never started appears on the date it was created, so it does not disappear from the calendar.

Things worth knowing:

• The calendar always shows all of your sessions. The search box, the state filters, and the tag filters apply to the List view only.
• On a keyboard, the arrow keys move between days (moving past the edge of the month turns the page), and Home and End jump to the start and end of the week.

Tagging Sessions

Tags are your own labels for a session — Onboarding, Client, Workshop, whatever helps you find it again. A session can carry any number of tags, and a tag you have used before is offered again, so the same label is spelled the same way each time.

To tag a session:

• When you create it — use Tags (optional) in the New session form.
• Afterwards — open the session, select the settings button in its title bar, and use Tags at the top of Session settings.

In either place, type a tag and press Enter, or pick an existing tag from the list that appears as you type. Select the × on a tag to remove it; Backspace in an empty field removes the last one.

Tags belong to a scope, like sessions do: your personal sessions share your personal tags, and a team's sessions share the team's tags.

To filter and search by tag, use the List view at one.vu.studio/ai/sessions:

• Your tags appear as chips after the state filters (All, Scheduled, Not started, Ended). Select one or more to show only sessions carrying any of the selected tags; select a chip again to clear it.
• The selected tags are kept in the page address, so a filtered list can be bookmarked or shared.
• Search sessions by name or focus… also matches tag names.
• Each session in the list shows its first two tags, plus a count of any others.

You can also ask Vū AI about sessions by tag — see Finding the session you mean.

The Session Console

Opening one of your own sessions gives you the owner's console. It is the same live session view a contributor or a cast wall sees — the immersive, full-screen picture of the conversation — with an extra switcher across the top for the surfaces only the session owner needs:

Tab	What it shows
Mind Map	The live concept graph: topics as nodes, sized by how often they come up, connected by how they relate.
Flow	The insights timeline (see below).
Devices / Breakouts	Who and what is contributing audio — see Devices and Breakout Rooms.
Dashboard	The generated insights dashboard (see below).

Alongside the switcher, the title bar carries Chat (ask Vū AI about this session, in a drawer that stays open across tabs), a download button for the Transcript as plain text or the Insights as structured data, Cast, and Settings. People who open the session through a join link, and walls you cast to, see the immersive view only — no switcher, no toolbar.

Switching Between Views

On your own session, the Insights tab holds every way of looking at the conversation. A view switcher floats at the bottom of the screen, just above the record bar:

View	What it shows
Mind Map	The live concept graph.
Flow	The insights timeline — see The Insights Timeline.
Cards	The insight board: the session's insights laid out as readable cards.
Dashboard	The generated insights dashboard — see The Insights Dashboard.
Report	The structured session report — see The Session Report.

The Insights / Transcript sidebar stays beside every view and can be collapsed. Devices (or Breakouts) is a separate tab at the top of the page.

A Shared session offers Mind Map, Flow, and Cards only. The dashboard and the report are saved to their owner's library, and a shared session has no single owner.

The Cards View

Cards shows the insight board, which is easier to read at a glance than the mind map. It is also offered wherever you pick a view for someone else:

• in the Cast to a wall dialog, under Opens in — see Casting a Session;
• when you share a device link from Add devices, under Opens in — see Adding Devices.

Unlike Mind map and Flow, which are chrome-less layouts for a wall, Cards keeps the insights and transcript sidebar alongside the board.

The Insights Dashboard

The Dashboard tab is a pass over the session's insights — not the transcript — that turns them into a readable set of cards: a short headline and summary, then metrics, rankings, timelines, comparisons, and topic clouds, chosen to fit whatever the session actually produced. Because it is built from insights rather than the transcript, it survives transcript deletion.

• It is generated on first open and kept with the session, so reopening it later shows the same dashboard rather than regenerating from scratch.
• Refresh regenerates it on demand — after more of the conversation has happened, or after you change how it is framed.
• Tune sets Viewing as (a role, which defaults from your own job title, so the dashboard leads with what that reader cares about) and an optional Focus ("I specifically want to know…"). Applying either regenerates the dashboard.
• Pin a card to keep it, and its format, through the next refresh; drag cards to reorder them.
• Export produces a PDF in a dark or a light theme (light is the one to use for printing on paper). Publish shares it as a link, and Edit with Vū AI opens it in a chat so you can keep shaping it by prompt.

For when the dashboard is generated, and what its Final and Interim labels mean, see When the Report and Dashboard Are Written.

To compare several sessions at once, tick two or more on the sessions list and select Combined report — you get the same Mind map and Insights pair across all of them.

The Insights Timeline

The Flow tab is an alternative to the mind map for watching a session unfold: insight cards drift across the screen as the conversation produces them, newest first, with a card's size, sharpness, and speed set by how much the session keeps returning to what it says. Each card is colour-coded by its axis — Signal, Theme, Tension, or Opportunity — and stamped with when it was captured.

It reads well on a wall or a shared screen, which is why it is also one of the views you can cast, share a join link into, or have a meeting agent screen-share back into a call. Where the mind map answers what is this conversation about, the timeline answers what has been surfacing, and in what order.

The Session Report

The Report view holds a structured written report of the session. It replaces the free-form white paper as the standard write-up of a session, and it is always laid out the same way:

1. Summary
2. Decisions
3. Next steps
4. Main themes
5. Open questions
6. What this is based on

A section with nothing in it is left out rather than printed empty — a meeting that decided nothing has no Decisions section. Summary and What this is based on always appear. The last section says how much of the conversation the report could draw on, and notes when the transcript was deleted.

Statements in the report carry the same claim labels as the insights (for example Decision made or AI inference), placed after the statement — see How Insights Are Labelled. The Main themes are the ones the Mostly about card identifies, so what the report emphasizes is what the conversation actually kept coming back to. The report is written in the session's Spoken language.

To open it, go to the session at one.vu.studio/ai/sessions, stay on the Insights tab, and select Report in the view switcher at the bottom of the screen.

When the Report and Dashboard Are Written

The report and the dashboard are both written once, when the session ends, from everything the session captured. That end-of-session version is free, and it is marked Final.

• While the session is still running, the Report view explains The report is written when the session ends and offers Write one now, with its token price on the button. The Dashboard view offers Build one now the same way. An early version covers the conversation so far and is marked Interim; the end-of-session version still runs afterwards and replaces it.
• Just after the session ends, the view shows Writing the final report for a minute or two.
• If the session had too little in it, the view says No report was written and offers Try again. Nothing is charged for a generation that comes back empty.

Shared sessions get neither a report nor a dashboard.

Working with the Report

Once there is a report, the header shows its title, a Final or Interim pill, and when it was generated. The toolbar offers:

• Open in a new tab (the arrow icon, far left) — opens the report as its own page.
• Tune — set Written for (who the report is for) and an optional Focus, then Apply to rewrite it. The next automatic rebuild uses the same settings.
• Edit with Vū AI — opens the report in a new Vū AI chat so you can keep shaping it by prompt.
• Publish — publish the report or share a public link to it.
• Export — export a PDF in Dark or Light. Light is best for printing on paper.
• Rebuild — write the report again from the session, at the token price shown on the button.

To rename the report, select the pencil beside its title.

To edit one section:

1. Hover over a section heading (or move keyboard focus to it) and select the pencil beside it (Edit the ... section).
2. Change the text. The heading stays as it is, and the rest of the report stays in view.
3. Select Save.

Each save, and each rename, is saved as a new version of the report. Editing text and renaming are free. Rebuild replaces hand-edited text, so it asks first if you have unsaved changes.

The report is kept as a microsite in your library, which is how it gets publishing, sharing, and version history. You can also ask Vū AI in chat for "the white paper" or "the writeup" of a session; it finds the report.

How Insights Are Labelled

Every insight has one of four axes, shown as a coloured badge: Signal, Themes, Tension, or Opportunity. Insights are not only a record of what was said. Vū Intelligence also points out things the room might not have raised itself:

• Tension includes risks and blind spots, including ones nobody mentioned.
• Opportunity includes suggested next steps, written as action items.
• Proactive suggestions can also appear under Signal and Themes where they fit.

These sit alongside the straightforward insights about what was discussed, which are still captured as before.

Many insights also carry a second, quieter label that says what kind of claim it is:

Label	Meaning
Agreed upon	The group converged on it and nobody pushed back.
Decision made	The group committed to doing something.
Recurring theme	The conversation kept coming back to it.
Someone's words	Close to what a participant actually said.
One person's view	Raised by one speaker and not picked up by anyone else.
Still open	Raised and discussed, but left unresolved.
Suggested action	A proposed next step nobody has agreed to yet.
Stated fact	A number, date, name, or concrete claim presented as fact.
Speculation	Hypothetical or exploratory — "what if we…".
AI inference	Vū Intelligence connected things nobody said outright.

The axis says what sort of thing an insight is; the kind says how solid it is. Many insights have no kind label: one is only added when it clearly applies.

The Mostly About Card

At the top of the insights sidebar, a Mostly about card names the three to five ideas the conversation kept coming back to. It is there in every view mode.

Each theme shows two counts: how many times it came up in the conversation, and how many other ideas it connects to in the mind map. Hover over a count to see which is which. The bar under each theme shows how it ranks against the top theme.

The card is worked out from the mind map, so it needs a conversation with some shape to it. It appears once the mind map has at least four ideas, and only ideas that came up more than once, or connect several others, count as themes. With breakout rooms, it follows the room selected in Filter by breakout room.

Connection Strength in the Mind Map

Connections in the mind map get stronger each time the conversation links the same two ideas again. A connection the room kept returning to is drawn thicker and more solid; one mentioned once is thin and faint. Strength is relative to the strongest connection in that session, so even a short session shows which links mattered most.

Low-Confidence Transcript Lines

A transcript line that may not have been heard accurately is marked with a small warning icon beside the speaker label. Hover over it (or focus it) to see This line may not be fully accurate. The icon is the only change — the line is not hidden or dimmed, and no score is shown.

Typed chat lines, and lines captured by the browser's own recognizer (the Just me setup), have no confidence reading, so they are never marked.

Very low-confidence lines stay in the transcript, where you can read and correct them, but are left out when insights, the mind map, and the report are built.

Editing the Transcript After a Session

Once a session has ended, its owner can correct the transcript in place:

1. Open the session at one.vu.studio/ai/sessions (Vū Studio icon → Vū AI → Vū Intelligence).
2. In the sidebar on the Insights tab, switch to Transcript.
3. Use the buttons at the right of any line:
   • Edit line — change the text, then Save (or Cancel; Esc also cancels).
   • Reassign speaker — choose Reassign to… and pick one of the session's existing speaker labels. This moves the line between labels already in the session; it does not name speakers.
   • Delete line — the line stays in place, struck through, with Restore beside it.

Edited lines show a pencil marker. Hover over it to see the original, for example Originally: "…" or Originally attributed to Speaker B.

Deleted lines can always be recovered. Select Show deleted above the transcript to list them as Deleted lines, then Restore a single line, or Restore all when there are several.

Things worth knowing:

• Editing needs an ended session. While a session is running or paused, the transcript shows End the session to enable editing.
• Only the session owner can edit, and only on the session page. Contributors, cast walls, and people who open a join link never see the editing controls.
• Shared sessions cannot be edited — the controls do not appear at all.
• Deleting the transcript removes deleted lines too. If the session is set to delete its transcript when it ends, or you delete it from Session settings, deleted lines go with the rest.
• Correcting the transcript does not rewrite insights or the report that were already generated.

Exporting a Session

The download button in the session's title bar offers three files:

Download	What you get
Transcript	Plain text (.txt). Each turn is time-stamped and labelled with its speaker. Chat lines are marked (chat), corrected lines (edited), and low-confidence lines [low confidence].
Insights	Structured data (.json).
Insights (CSV)	A spreadsheet (.csv) with each insight's axis, confidence, text, when it was created, and the device that captured it.

Exports always reflect the corrected transcript: edits are included and deleted lines are left out.

Chat with a session

Any session is chattable through Vū AI, just like the rest of your workspace — ask it to summarize the discussion, pull out decisions and action items, or draft a recap.

Questions are answered from the whole session, however long it ran — the full transcript, every insight, the concept graph, and the session's running summary. Ask for the action items from a two-hour meeting and you get all of them, not the most recent handful. Because the transcript is attributed and time-stamped, answers can point at the moment something happened — "around 08:31 Speaker B pushed back on the timeline" — rather than paraphrasing the discussion as a whole.

For the same reason, Vū AI can line an insight up against the part of the conversation that produced it, so "why did it flag that?" is a question it can actually answer.

There are two places to do it:

• In the session — select Chat in the session's title bar. The drawer stays open as you move between the Mind Map, Flow, Devices, and Dashboard tabs.
• In the main Vū AI chat — at one.vu.studio/ai (Vū Studio icon → Vū AI), just name the session you mean.

Finding the session you mean

You do not need the session's exact title. Describe the meeting the way you would to a colleague and Vū AI works out which one you mean — for example:

• "the meeting last Tuesday where we talked about the warehouse rollout"
• "the session about tiered pricing"
• "what did we decide about the boardroom install?"

It matches your description against each session's title, its focus, its running summary, and the text of the insights it captured — which is what lets a topical description work at all, since a session's title is often taken from a calendar invite and shares almost no words with how you refer to the meeting afterwards. A title match counts for the most; a topic mentioned in passing inside an insight counts for the least.

Time words in your description are read as a date range rather than as search terms, so "last Tuesday" narrows when the session was created instead of hunting for the word "Tuesday" in your titles.

You can also find sessions by their tags — see Tagging Sessions. Ask "What session tags do I have?" for a list of your tags with how many sessions carry each, or name a tag in your request:

• "Show me my onboarding sessions" — sessions carrying that tag.
• "Get my client or workshop sessions" — sessions with either tag.
• "Which sessions are tagged both client and workshop?" — only sessions carrying every tag you name.
• "Summarize my latest workshop session" — tags combine with the rest of the request, including dates.

Tag names match exactly, ignoring capitals and extra spaces, so work does not match a tag called homework. If no tag matches what you asked for, Vū AI says so, or offers the closest tags you do have, rather than guessing. Vū AI uses the tags of the scope you are working in: your personal tags, or the active team's.

Two things it does rather than guess:

• If several sessions match your description equally well — usually the same topic across a run of meetings — it lists the candidates and asks which one you meant, with a line from each session's own insights so you can tell them apart in your own terms.
• If nothing matches — it offers your most recent sessions to pick from instead of reporting that no session was found.

You only ever see sessions you own or were added to.

Opening a Session Microsite in the Editor

Ask Vū AI inside a session to build a microsite — a recap page, a summary deck, a whitepaper — and the finished page appears as a card in the session's chat. The card carries an Open editor button, which takes you to the full microsite editor in the main Vū AI chat, on the same conversation, with that microsite already open. (The button only appears if you can edit the microsite; someone with view-only access to it does not see it.)

The editor lives outside the session console, so if this device is currently recording you are asked what to do first. The dialog is headed You're still recording and offers:

Choice	What happens
Open in a new tab	The editor opens in a second tab and this device keeps recording. Use this one unless you have a reason not to.
Stop and open	This device leaves the session console, which stops its capture, and the editor opens in place.
Cancel	Nothing happens; you stay in the session.

Only this device's capture is affected. The session itself stays live, and any other joined devices or meeting agents keep recording — see Devices and Breakout Rooms.

If you are not recording, Open editor simply takes you to the editor with no prompt.

Attendee-aware follow-ups

Because Vū Intelligence captures attendees (with their emails when available), you can act on them afterward. For example, ask Vū AI to generate a recap, whitepaper, or microsite and email the link to everyone who attended. Follow-ups that need email addresses only reach attendees whose email was captured.

note

For a Shared session (one involving several people from your organization), follow-up actions are handled by asking Vū AI after the meeting rather than through a pre-set end-of-session automation. Anyone with access to the session can ask Vū AI to act on it.