> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kair.is/llms.txt
> Use this file to discover all available pages before exploring further.

# Recording a Session

> Step-by-step: from creating a session to getting a transcript.

## Before you start

* You need a moderator or admin account
* The workspace must already exist (ask an admin to create one if needed)
* Have the Kair Voice web app open at `kair.is`

## Step 1 — Create a session

1. Go to your workspace dashboard
2. Click **New Session**
3. Give it a title (e.g. "Town Hall — 27 May 2026")
4. Optionally set the planned headcount
5. Click **Create**

You'll land on the session page.

## Step 2 — Start recording

Click **Start Recording**. The button turns red and a timer starts. All participants in the room are now being recorded.

<Info>
  If someone else is already recording the same session, you'll see a "Recording in progress — take over?" prompt. Only click take-over if the original recorder has disconnected.
</Info>

## Step 3 — Monitor live transcription

As audio chunks are uploaded (every few seconds), partial transcripts appear in the moderator view in near real-time. These are preliminary — the final transcript (with speaker labels) is produced after you stop recording.

## Step 4 — Pause / resume (optional)

Click **Pause** to temporarily halt recording without ending the session. Click **Resume** to continue. The same session ID is maintained — chunks from before and after the pause are stitched together.

## Step 5 — Stop recording

Click **Stop Recording**. The session moves to `finalising` state:

1. All chunks are stitched into one audio file
2. A final transcription job runs on the full audio (with diarisation if enabled)
3. Speaker labels (`S1`, `S2`, etc.) are assigned to segments

This takes a few minutes depending on session length and the transcription backend.

## Step 6 — Request a summary

Once transcription is complete, click **Generate Summary** on the session page. Summaries are not automatic — you trigger them when ready. The AI produces a concise summary of what was discussed.

## Step 7 — Review the transcript

The full transcript is available under the **Transcript** tab. Segments show speaker labels and timestamps. You can see who said what and when.

***

## Troubleshooting

| Problem                       | What to do                                                         |
| ----------------------------- | ------------------------------------------------------------------ |
| Recording button greyed out   | Someone else is recording — see the take-over prompt               |
| Chunks failing                | Check Admin → Jobs for failed chunk jobs; re-run individually      |
| No speaker labels             | Diarisation may be disabled — check Admin → System → Transcription |
| Summary not appearing         | Click "Generate Summary" — it's not automatic                      |
| Session stuck in `finalising` | Check Admin → Jobs for a failed final transcription job            |