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

# Semble

> Private video meetings with an independent AI participant — talk to Kai by name and get answers out loud, grounded in your documents.

Semble is a **video meeting in your browser** with one extra participant whenever you want it:
**Kai**, an independent voice AI that belongs to no one and is equal to everyone in the room.
Say its name and ask — Kai answers out loud and in text, with citations from the documents
shared with the meeting, keeps a live transcript, and writes a neutral summary afterwards.

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/semble/quickstart">
    Start or schedule a meeting, invite people, and join as a guest.
  </Card>

  <Card title="Talking to Kai" icon="sparkles" href="/semble/talking-to-kai">
    Turn Kai on, ask by voice, and get answers everyone hears.
  </Card>

  <Card title="Shared documents" icon="file-lines" href="/semble/documents">
    Give Kai the material the meeting is about — answers cite it.
  </Card>

  <Card title="Transcript & summary" icon="list-check" href="/semble/transcript-and-summary">
    A live, attributed transcript and a neutral group summary.
  </Card>

  <Card title="Recording" icon="circle-dot" href="/semble/recording">
    Optional audio recording — everyone is told, nothing is silent.
  </Card>

  <Card title="Hosting" icon="shield-halved" href="/semble/hosting">
    Lock the meeting, remove a guest, and the free-meeting limits.
  </Card>

  <Card title="Credits" icon="coins" href="/semble/credits">
    Meeting is free; prepaid credits pay for Kai's time and recordings.
  </Card>
</CardGroup>

## The idea in one minute

<Steps>
  <Step title="Start a room">
    Sign in with an email code, start an instant meeting or schedule one, and share the link.
    Guests join from the link with just a name — no account, no install.
  </Step>

  <Step title="Bring Kai in">
    When the conversation needs it, the host turns Kai on. Kai joins as a participant tile and
    starts listening.
  </Step>

  <Step title="Ask by name">
    Anyone in the room — guests included — says "Kai, …" and asks — in English, Russian, and many other languages — or types the
    question. Kai answers out loud for everyone, with the same answer and its sources shown in text.
  </Step>

  <Step title="Walk away with the record">
    After the call the host has the attributed transcript, a neutral summary of key points,
    decisions, open questions and action items — and, if they recorded it, the audio.
  </Step>
</Steps>

<Note>
  Kai is nobody's assistant. It treats every participant the same, answers only when addressed,
  and never interrupts — if someone starts talking, it stops.
</Note>