Preview — payments disabledYou can create an account and use the site, but checkout is switched off and no subscription can be started here.
Skip to main content
Xuula

New in 1.9.0

Meetings: follow the conversation, keep the record

Start a meeting recording and the app captures your microphone and the Windows system audio — the other side of the call — into one recording, while a live transcript builds in front of you. When you stop, the recording, transcript, translation and notes are saved together as one conversation.

An illustration of the live Meeting Workspace: a speaker-separated transcript with a translation beside it, a recording indicator, a microphone mute switch, and the Project the meeting will be saved to.

How a meeting flows

  1. 1

    Start from the mini panel

    A draggable, collapsible always-on-top panel with its own hotkey. Pause and resume, watch elapsed time, mute your microphone with one switch, and control the caption windows without opening anything else.

  2. 2

    Follow the live transcript

    Words appear progressively as they are recognised, with speakers separated and labelled. Always-on-top Original and Translation caption windows can be shown, moved, restyled, closed and reopened from the mini panel — without stopping the recording.

  3. 3

    Ask the AI while it runs

    The Meeting Workspace keeps the live transcript, translation and your private notes together. On-demand AI actions — Catch Up, Summarize So Far, Action Items, Suggested Questions or a free-form question — run only when you ask, on the transcript text.

  4. 4

    Stop, and it saves to a Project

    Choose “Saving to” directly in the live workspace. A visible saving state runs, and the recording, final transcript, translation, notes, speaker names and word timestamps land in the Project you picked as one conversation.

Both sides of the call, one recording

Microphone and Windows loopback audio are normalised into one synchronised stream with echo suppression, saved as a single audio file. One recognition session produces a chronological, speaker-aware transcript — you are not paying for two.

Speakers you can name

Detected speakers are separated automatically and labelled. You can rename yourself, the meeting audio, and each detected speaker for the current session, and the names are kept with the saved conversation. Speaker separation is acoustic — it is not identity verification.

Translation beside the original

Optional live translation shows the original and a selected target language side by side, both during the meeting and in the saved conversation. The original transcript is always kept.

Pauses are not sentence breaks

Local cleanup joins runs of fragments caused by thinking pauses instead of splitting one thought into pieces, and the saved transcript uses only confirmed recognition tokens. An optional final AI polish is applied when configured — with the local result as the fallback.

After the meeting

A saved conversation carries persistent, on-demand AI tabs: Summary, Key Points, To-do and Speaker Quotes — generated one at a time or all four in sequence, and kept with the conversation so you never regenerate them just to reread them.

AI output can contain mistakes. Summaries, to-dos and answers are drafts to review, not records to rely on blindly.

What Meetings is not

  • It does not join Zoom, Teams or Google Meet as a bot. It captures the audio your Windows PC plays and your microphone, locally — which works with any app that plays sound, and needs no meeting-platform account.
  • It cannot capture audio that Windows will not expose — DRM-protected paths or some exclusive-mode audio setups.
  • Recording people may require their notice or consent depending on where you and they are. That responsibility is yours; the app makes no legal determination for you.
Download for Windows

2,000 words free every week. No credit card.