Meetings

Record or upload meetings, transcribe them automatically with speaker separation, and turn them into minutes, action items and other documents.

The Meetings module turns spoken meetings into searchable, structured text. You record a meeting directly in your browser, upload an existing audio or video file, or (where enabled) send a recording bot into your online meeting. Kasimir transcribes the recording with automatic speaker separation, then produces full minutes with a summary and action items, and lets you generate further documents such as follow-up emails or decision lists at the click of a button. All processing runs GDPR-compliant on Kasimir's own infrastructure – no US transcription service is used.

What you can do with this module

• Record a meeting live in the browser (microphone).

• Upload an existing recording (audio or video).

• Invite a recording bot into an online meeting (Teams, Zoom, Google Meet) – if enabled for your workspace.

• Read the transcript with timestamps and speakers and jump along in the audio.

• Give the detected speakers (SPEAKER_00, SPEAKER_01 …) real names.

• Generate various documents from the transcript (minutes, task list, follow-up email and more) and export them as Markdown, Word or PDF.

• Add the transcript to the knowledge base so chat can use it.

Capturing a meeting

On the Meetings page, the top section Step 01 – Join or upload offers all three ways to start a recording.

Option A – Invite a bot to the meeting

1. Paste the invite link of your online meeting into the field (https://… Teams / Zoom / Google Meet).

2. Click Invite. The bot joins the meeting as a participant (Kasimir Notetaker) and records it.

3. A live status appears below the form, moving through Requesting bot…, Bot joining…, Bot in call, Bot recording…, Processing recording… to Done.

4. Once the status is Done, the Open meeting button takes you straight to the finished recording. Use Dismiss to close a completed or failed status message.

Supported platforms are Teams, Zoom and Google Meet. Other links are rejected with an error.

Option B – Upload a recording

1. Click the upload tile (Upload meeting) and choose a file – or drag it onto the area.

2. The upload starts automatically; progress is shown as a percentage.

Supported formats and limits:

Option C – Record live in the browser

The recorder sits next to the upload tile.

1. Click Record and grant microphone access.

2. While recording you see a waveform visualization, a blinking record dot and the running duration. The screen stays awake (no sleep) while recording.

3. Use Stop & upload to end the recording; it is uploaded and processed automatically.

What happens after capture

Every recording automatically goes through several phases in the background. You don't have to wait – the list and detail page refresh on their own.

1. Pending – the recording is queued.

2. Transcribing… – speech is converted to text and speakers are detected.

3. Summarising… – minutes, summary and action items are generated.

4. Ready – the recording is fully analysed. (On problems: Failed with an error message.)

On the detail page an animated progress indicator shows the steps Upload → Transcribe → Summarize.

The meeting detail view

Clicking an entry in the list opens the meeting. At the top you see the title, date, language code and the number of detected speakers. Use the pencil icon to rename the title and the trash icon to delete the meeting (with a confirmation).

Below sits the audio player, which is linked to the transcript, plus three tabs.

"Summary" tab

Shows the automatically generated Markdown minutes (with sections such as participants, topics, decisions) and, beneath it, the suggested actions as a task list with owner and due date where identifiable.

"Protocol" tab

This holds the full transcript organised by speaker and timestamp.

• Each line shows the timestamp on the left and the colour-coded speaker. Clicking a line jumps the audio to that point; the line currently being spoken is highlighted and scrolled into view.

• At the top you can name the detected speakers: type the real name into the field next to SPEAKER_00 etc. Use Listen to play a short voice sample of that speaker to identify them. Then click Save.

"Templates" tab

Here you generate ready-made documents from the transcript. Generate output opens a list of templates.

Available templates:

1. Choose a template. For Custom template a text field appears where you describe in your own words what should be generated (up to 2000 characters), then click Generate.

2. The new output appears as a card with the status generating … and fills in automatically once finished.

3. On a finished card you have: Copy (to clipboard), Export and Delete.

Export formats: Markdown (.md), Word (.docx) and PDF (.pdf).

Add to knowledge base

At the bottom of the detail view you find Add to knowledge base. This stores the transcript as a company-wide document and makes it searchable in chat (RAG). The button then changes to Already in knowledge base.

Visibility and organisation

• Private by default: A meeting is only visible to you (its creator) – marked with the Private badge. Others in the company can't see it.

• Project-scoped: If a meeting is assigned to a project, it becomes visible to all members of that project.

• Filter: Above the list you filter using the All, General (no project) buttons and one button per project.

Tips & limits at a glance

• Maximum file size: 200 MB; supported formats MP3, MP4, M4A, WAV, WebM, OGG.

• Bot: at most one active bot per workspace, only Teams/Zoom/Google Meet; may not be enabled (Coming soon).

• Language: transcription is optimised for German.

• Excel export is visible in the menu but currently non-functional – use Markdown, Word or PDF.

• Recovery: interrupted browser recordings can be finalised afterwards.

• If you leave the page during processing, nothing is lost – the analysis keeps running in the background.