Integrate

One iframe to show her, one call to make her talk.

Both of these work as they are. Replace the token and the room name and nothing else needs to change.

01 Put her on the page

Embed

450 × 600, rounded, transparent behind — she takes your own background. Keep allow="autoplay" or the browser blocks her voice and she mouths silently.

<iframe
  src="https://avatar.threadlane.in/widget.html"
  width="450" height="600"
  style="border:0;background:transparent"
  allow="autoplay"></iframe>

02 Let your agent drive her

MCP

Streamable HTTP. Point any MCP client at it and she gains speak, gesture and status.

{
  "mcpServers": {
    "avatar-ai": {
      "type": "http",
      "url": "https://avatar.threadlane.in/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Claude Code, in one line:

claude mcp add --transport http avatar-ai https://avatar.threadlane.in/mcp \
  --header "Authorization: Bearer YOUR_TOKEN"

03 Or call it directly

HTTP

One round trip: she writes the line if you do not supply one, renders it, and returns the audio. The response body is the mp3.

curl -X POST https://avatar.threadlane.in/api/v1/say \
  -H "Authorization: Bearer $BRIANA_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"prompt":"Introduce yourself to a new visitor."}' \
  --output briana.mp3

The full reference — every route, every gesture, rooms, limits and errors — is in the documentation. Need a token, or a custom avatar? connect@threadlane.in.