Skip to main content

Conversational Intelligence

Ask questions while your Voilo Assistant is in the call and get answers back in milliseconds. The Threads API gives every Meeting a Slack‑style chat lane – perfect for side‑bar Q&A, action‑items, or "hey bot, do the thing" tool calls.

How It Works ⚡️​

  1. Post a question to POST /v1/meetings/{meeting_id}/threads/{id}/messages.
  2. Assistant reads the transcript, crafts a response (or detects a tool request).
  3. Stream replies in real‑time over WebSocket GET /v1/meetings/{meeting_id}/threads/{id}/stream.

Under the hood each chat Message is JSON with role, content, timestamps, and (for transcript segments) a segment_offset. See full schema in Messages API.

Real‑World Hook 🕴️ Sales‑Call Objection Tracker​

Picture a high‑stakes demo. Mid‑call the prospect drops a classic: “Price is my biggest blocker.” Your sales rep triggers a chat shortcut:

curl -X POST \
  https://api.voilo.io/meetings/mtg_123/threads/main/messages \
  -H "Authorization: Bearer $VOILO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "role": "user",
    "content": "What objections have we heard so far?"
  }'

Two seconds later the Assistant streams back:

{
  "role": "assistant",
  "content": "Top objections so far:
    ① Budget (price too high),
    ② Timeline (Q3 start),
    ③ Security review needed."
}

You’ve armed the rep with instant ammo to tackle concerns before the meeting ends – while the buying temperature is hot.

Tips & Gotchas 🧰​

  • One thread per Meeting is created by default. Create extra via POST …/threads if you need side‑channels.
  • Latency ≈ 500 ms from user message → first assistant token (depends on LLM).

See also​