Live Transcribe
Live Transcribe captures a session as it happens and turns it into a speaker-attributed transcript, an AI summary, insights, and — if you want them — recap drafts. It runs in the browser, so there’s no app to install and no hardware beyond a microphone or line input.
Video transcript
Live Transcribe turns a session in a room into a transcript, a summary and a recap while it’s still happening. It’s under Marketing, and the sidebar calls it Transcribe. Overview, Recaps, Attendee View, Invites, Settings. Two ways to run it. Your own team, from a session’s console right here in Sessionboard. Or an AV crew, from the Live Transcribe Console with a magic link and no login. Every captured session lands down here, with badges for transcript, summary and insights.
Settings, before the event. The model, General or Medical, and Medical earns its keep on a clinical program. Which session statuses may be transcribed, so a closed-door session never appears in the picker. Keyterms, your drug names, acronyms and speaker names. The AI outputs, summaries, insights and auto-generated recaps, drafts only. And under Advanced, Post-session recovery, which re-transcribes from the recording once the session ends and fills in anything the live pass missed. Room Console and Mission Control are for multi-room shows.
Before show day, each room needs a few things. A laptop with Chrome. A wired internet connection, not the venue Wi-Fi. The soundboard or microphone feed running into that laptop’s audio input. And one person assigned to start and stop each session. That person needs event access, either a team member invite, or a Live Transcribe Console invite link from the Invites tab.
Here’s a session on the day, from its console. Pick the audio input. A line from the soundboard beats a laptop microphone in a big room. The meter under the picker moves when the feed is live, and Test input runs a short transcription that isn’t saved. Then Start Transcription. Nothing starts by itself. Someone presses this once, for each session.
And it’s running. Recording live, the fragment count ticking up, and the audio saving as it goes. The Live Transcript fills in as people speak, a line at a time, each one stamped with who was talking. Speaker one, speaker two. Those voices get real names afterwards. And under Recordings, the take is saving to disk as it goes, so there’s always audio to fall back on.
When the session ends, Stop Transcription. The take finishes writing, and if auto-recaps are on, a recap draft is queued. With Post-session recovery on, the recording is re-transcribed in the background, so a dropped connection or a late start isn’t lost. Transcript and recap takes you to everything this session produced.
Now the public side. Attendee View. The page is private until you flip Publish attendee view. That makes the transcripts and summaries of every captured session readable to anyone with the link, so read that line before you switch it on. Share and embed gives you the link, a QR code for signage, and an embed for your own site. I’ll publish it, and let a session run in the room so you can see it live.
This is what an attendee sees at that link. Your branding, a searchable schedule split into Upcoming and Past, as a grid, by day or by room, and Sessions, Speakers and Media across the top. And because a session is being captured right now, a Live now band across the top, with Watch live.
Watch live opens the live stage. The same captions we saw in the console, streaming to anyone in the room or at home. Insights as they build, Details for the schedule. Underneath, reactions anyone can tap, and a live discussion. I’ll send one. Those reactions and comments are public, the moment they’re sent.
For a session that’s over, the attendee’s page carries whatever exists for it. Here’s one from earlier. The speakers, the AI recap, highlights, the attachments, and the full transcript, with a PDF and a share image to take away. Only what was captured is here, and only while the view is published.
On your side, the same session under Media. Each line, a timestamp and who said it. Search the transcript, filter by speaker with these chips, and hover a line to correct it in place. Tag names a diarized voice, a session participant or a generic tag like Audience Member. Recaps, exports and the attendee page use the corrected text from then on.
Recaps. One event report, and a recap per captured session, each drafted from the real transcripts. Executive summary, key takeaways, topics, a LinkedIn caption, each a card you can edit. Everything lands as a Draft. Approve it, then Publish puts it on the attendee view. Nothing publishes on its own. Export gives you a PDF, a LinkedIn carousel or a share card.
So that’s Live Transcribe. Set it up once, someone presses start in the room, and the captions, the transcript, the recap and the attendee page follow. The Help Center has this under Marketing, Live Transcribe.
Why use it
Section titled “Why use it”What a speaker says usually leaves the room with the people who heard it. Live Transcribe listens to the room’s audio — a soundboard feed or a microphone — and writes it down line by line while the session runs, each line stamped with who was speaking, so nobody has to take notes. Once you publish the attendee view, attendees can follow the live captions in the attendee hub. When you stop, the transcript is ready to search and correct, an AI summary and insights follow shortly, and, if auto-recaps are on, a recap draft is queued for your approval. Keyterms and the Medical model keep drug names, acronyms and speaker names accurate on clinical programs.
Two ways to run it
Section titled “Two ways to run it”Which one you use depends on how your event is staffed.
| Self-serve | AV crew | |
|---|---|---|
| Who runs it | Your team, from the admin app | AV professionals, from the Console |
| Where | Marketing → Transcribe | Live Transcribe Console at a separate URL |
| Login | Normal Sessionboard login | Magic link, no account |
| Best for | A handful of rooms you’re staffing yourself | Multi-room events with a dedicated AV team |
Both produce identical output. You can mix them across an event.
What each room needs
Section titled “What each room needs”Sessionboard runs in the browser, so nothing is installed on the room laptops. Plan on this per conference room:
- A laptop with Chrome. It stays in the room for the day and runs the Console.
- A wired internet connection. Live transcription streams audio continuously; venue Wi-Fi drops cause gaps a wired line does not.
- Audio input connected to the laptop. A line from the soundboard into the laptop’s audio input. A laptop microphone works in a small room, but a board feed is what makes a large room accurate.
- One person assigned to start and stop each session. Transcription never starts by itself. Name who presses Start Transcription and Stop Transcription in each room before show day.
Rooms that would rather run a desktop app than a browser tab can install Ready Room, the desktop version of the Console. The Download Ready Room card on the Live Transcribe page (and at the top of the Console home) offers the installer for your operating system first — macOS, Windows, or Windows on ARM — with the others and a checksums link beside it, and shows the version. The links are made for you when you open the page and work for 15 minutes, so download from the card rather than saving the link; reopen the page for a fresh one. The card shows only while a release is published.
The people running the rooms need access to the event: either a Sessionboard login (invite them as event team members) or a Live Transcribe Console invite link, which needs no account (see Give your crew access).
Set up before the event
Section titled “Set up before the event”Go to Marketing → Transcribe → Settings.
- Sessions available to transcribe — which sessions appear in the Live Transcribe Console and the Capture session picker, and which ones capture can start on. Choose All sessions, or Only sessions with selected statuses and tick the standard and custom statuses to include; at least one must stay ticked. New events start with Accepted only. Events that were already set up keep showing all sessions until you change this. With a status filter on, withdrawn sessions are always left out. The attendee-facing session list is not affected.
- Transcription model — General or Medical. Medical is worth selecting for clinical and scientific programs.
- Multilingual live captions — turn this on when a room mixes languages on the same floor feed (for example English and French). The live model detects the language per utterance. Leave it off for English-only rooms. If a room must stay English-only live, feed the interpretation channel instead of the floor mix. When this is on, live capture and post-session recovery use General (Nova-3), because Medical is English-only. Medical still applies when the setting is off.
- Keyterms — product names, acronyms, and speaker names the model would otherwise miss.
- AI summaries and AI insights — generate a summary and pull quotes and topics per session.
- Auto-generate recaps — drafts a recap when a session stops. Drafts only; nothing publishes on its own.
- Translation — stores translated transcripts for admin use. Attendees pick their language in the attendee hub Settings. That picker is separate from this toggle.
- Post-session recovery — re-transcribes from the recording afterwards, which cleans up anything the live pass missed. Re-transcribing regenerates every line, so wait for it to finish before correcting transcript lines by hand — manual edits are discarded by a re-transcription.
- Speaker tagging — manage the Generic participant tags (Audience Member, MC / Emcee, Staff Member, Unknown Speaker, and any you add) used to label voices that are not session participants. See Tag speakers.
- Room Console and Mission Control — for multi-room AV operations. See Live Transcribe Console.
Capture a session
Section titled “Capture a session”Open the session
Go to Marketing → Transcribe and open the Console, then pick the session from the room’s schedule. Its session console is where the capture runs.
Choose your input
Select the microphone or line input. A feed from the soundboard gives far better results than a laptop microphone in a large room. The level meter under the picker moves when the feed is live. Click Test input to run a short transcription that is not saved to the session.
Start
Click Start Transcription. The transcript streams in live under Live Transcript.
Stop
Click Stop Transcription when the session ends. If auto-recaps are on, a recap draft is queued. In the Live Transcribe Console, use Next session in the Up next panel to move straight on to the next session in that room.
After the session
Section titled “After the session”The session appears under Captured sessions with badges for Transcript, Summary, Insights, and Podcast. Opening it takes you to the session profile in Media, where you can tag speakers — session participants, or generic tags such as Audience Member and MC / Emcee for voices that are not contacts — correct the transcript line by line — including changing who said a single line, splitting a line, and adding or removing lines — and pull quotes and clips.
Recap drafts land in Marketing → Transcribe → Recaps. Generating one takes a minute or two, and nothing publishes without you.
Consent and sensitive sessions
Section titled “Consent and sensitive sessions”Decide before the event which sessions may be captured, and keep the rest out of the pickers rather than deleting transcripts afterwards. Two controls do this:
- Sessions available to transcribe in Marketing → Transcribe → Settings. Choose Only sessions with selected statuses and give sessions that must not be recorded — licensed content, closed-door discussions — a status you leave unticked. They then do not appear in the Capture session picker or in the Live Transcribe Console, and capture can’t start on them from a direct link or a Ready Room station’s saved copy either: the panel says the session is not available to transcribe. A capture that is already running when you change the setting keeps going until it is stopped.
- Nothing is captured until someone clicks Start Transcription. For a one-off, simply do not start the session; the AV crew’s Live Transcribe Console shows the same list you configured.
The life of a captured session
Section titled “The life of a captured session”A capture moves through four stages, and each has one place to look:
- Capturing. From Start Transcription to Stop Transcription, the live transcript fills in and the audio saves as it goes. See Capture a session.
- Finishing. After you stop, the take finishes writing. With Post-session recovery on, the recording is re-transcribed to fix anything the live pass missed, and a recap draft is queued if auto-recaps are on.
- Captured. The session sits under Marketing → Transcribe → Captured sessions and as a Live session card in Marketing → Media → Library. Edit speakers and lines, pull quotes and clips, and send recaps from here.
- Deleted. A capture you no longer want, including test captures, is deleted from one of the places below. There is no archive step for live captures.
Delete a captured session
Section titled “Delete a captured session”If a session was captured by mistake, or you ran test captures before the event, delete the capture. There are three places to do it:
- Marketing → Media → Library, in the card layout: the ⋯ menu on the session’s Live session card, then Delete. The list layout shows the same session as a row without that menu. In the list layout, open the session and use the option below.
- The live session’s own page: Actions → Delete live session. You land back in the Media library afterwards.
- Marketing → Transcribe → Captured sessions: hover the row and click Delete.
The confirmation names the session. Deleting permanently removes everything the capture produced: the transcript, the saved audio, the AI summary, insights, translations, highlights, comments and the podcast recap. It disappears from Media, from Captured sessions and from the attendee view. The session itself stays on your agenda with its time, room and speakers, so you can capture it again later.
Do I need special hardware?
Nothing to install. Each room needs a laptop running Chrome on a wired internet connection, an audio input into that laptop (a soundboard feed gives the best quality), and a person assigned to start and stop each session. See What each room needs.
How long until I can use the content?
The transcript is live as the session runs. Summaries and insights follow shortly after you stop. Recap generation takes a minute or two.
What if the transcript missed the beginning?
Turn on Post-session recovery in Settings. It re-transcribes from the recording afterwards and grooms the result. If Multilingual live captions is on, recovery uses the same mixed-language model so French and Portuguese stay labeled per utterance. Any re-transcription — recovery included — regenerates every line and discards transcript lines you edited by hand, so make manual corrections after it has run.
Can attendees follow along?
Yes, through the attendee hub. Configure it under the Attendee View tab. Attendees pick their language in hub Settings — live captions, replay, and the session transcript follow that choice, including when the floor mixes languages.
What if speakers switch languages in the same room?
Turn on Multilingual live captions in Marketing → Transcribe → Settings and capture the floor mix (stage mics), not the interpreter booth. Speaker labels stay on the people on stage. Use the interpretation channel only if that room must be English-only live; then the speaker label is the interpreter.
Why don't I see the Invites or Console options?
Those appear for enterprise AV setups. Events on the simplified shell run capture entirely inside the admin app.
Related
Section titled “Related”- Live Transcribe Console — the AV crew’s console
- Media library — where transcripts and recordings live
- Content plan — turn a captured session into a month of content