Meetings

A meeting is a live AI conversation session. Creating one spins up a Daily.co video room with your AI persona ready to talk and returns the join URL.

Every account is auto-seeded with two ready-to-use scenarios — Plus One and Requirement Gathering. List them with GET /scenarios and look for the default_kind field. No need to build a scenario before your first meeting.

Create a meeting

The minimum you need is a scenario_id and a person_name.

curl -X POST https://api.waterr.ai/v1/meetings \
  -H "Authorization: Bearer wai_<your_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "person_name": "John Doe",
    "scenario_id": "scenario-uuid"
  }'

string

required

Name of the participant joining the session. The agent greets and addresses them by it, and it labels the call in transcripts, analyses and memory. If you don’t know who will join yet — say you are generating a link to send out — send "Guest".

string

required

UUID of the scenario this meeting runs.

string

Override the persona’s default voice for this session. See Voices.

string

regular (default) or instant.

string

Session-specific context injected into the AI’s prompt without changing the scenario. Use this to personalize: “Candidate has 8 years at Google, focus on ML infrastructure.”

string

The participant’s email. Fills {{USER_EMAIL}} / {userEmail} in the scenario’s prompt and welcome message.

string

Title for this call in meeting lists and analyses. Defaults to the scenario name. Max 200 characters.

integer

Length of this call, 1–240. Anyone can shorten a call; only your own workspace can run its scenarios past their saved length. Your plan’s call cap still applies.

Override the agent for one call

On your own workspace’s scenarios you can change the agent for a single call. The saved scenario is never touched — the next call runs it as saved. Sending any of these fields for another workspace’s scenario returns 403.

curl -X POST https://api.waterr.ai/v1/meetings \
  -H "Authorization: Bearer wai_<your_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "person_name": "Priya Shah",
    "email": "[email protected]",
    "scenario_id": "scenario-uuid",
    "system_prompt": "You are running a 10-minute churn interview. Ask why they cancelled, what would bring them back, and nothing else.",
    "welcome_message": "Hi Priya, thanks for making time — this will take ten minutes.",
    "language": "English",
    "controls": { "allow_recording": true, "end_call_enabled": true, "allow_web_search": false }
  }'

string

Replaces the scenario’s prompt for this call. The persona, voice, knowledge base, tools and every other scenario setting still apply, and note_context is still added on top. Plain text or a POML document, up to 64 KB (413 above that). Use note_context instead when you only want to add to the scenario’s prompt.

string

The agent’s first line for this call.

string

The call’s language, e.g. "English" or "Hindi". The voice follows it.

string

The name the agent goes by in this call.

object

Overrides the scenario’s meeting settings for this call. Only the keys you send change; unknown keys are rejected with 400. A capability your plan doesn’t include is switched off for the call, and the call is priced with the capabilities it actually runs.

Show controls

boolean

Record the call.

boolean

Keep a transcript.

boolean

Show the in-room chat.

Let the participant share their screen.

boolean

Let the participant turn on their camera.

boolean

Let the participant unmute.

boolean

Hold participants in a waiting room.

boolean

Admit waiting participants automatically.

integer | null

1–100, or null for no limit.

boolean

Let the agent end the call itself.

boolean

The agent sees the participant’s camera.

The agent sees a shared screen.

boolean

Use the vision model on what the agent sees.

boolean

The agent can search the web.

boolean

The agent can hand hard questions to a slower reasoning model.

boolean

Read and write memory for this participant.

string[] | null

Up to 10 start pages for the live browser demo; null turns it off.

object

A bring-your-own-agent endpoint for this call: { "base_url", "api_key", "model" }. Own workspace only.

Response

{
  "success": true,
  "data": {
    "id": "meeting-uuid",
    "daily_meeting_url": "https://waterr-xx.daily.co/abc123",
    "room_token": "eyJhbGciOiJIUzI1NiIs...",
    "meeting_url": "https://waterr.ai/meeting/meeting-uuid",
    "status": "created"
  },
  "session_minutes_info": {
    "estimatedMinutes": 25,
    "remainingMinutes": 75
  }
}

Field	Use
`meeting_url`	WaterrAI-hosted page — easiest way to drop a user in.
`daily_meeting_url`	Raw Daily.co room URL — for embeds or SDK joins.
`room_token`	Auth token for SDK joins.
`id`	Use this to fetch post-meeting data.

Joining the meeting

Hosted page — redirect the participant to meeting_url. Iframe embed:

<iframe
  src="DAILY_MEETING_URL?t=ROOM_TOKEN"
  allow="camera; microphone; autoplay; display-capture"
  style="width: 100%; height: 600px; border: none;"
></iframe>

Daily.co SDK for full UI control:

import DailyIframe from '@daily-co/daily-js';
const call = DailyIframe.createCallObject();
await call.join({ url: data.daily_meeting_url, token: data.room_token });

End-user meetings (no JWT)

For public or embedded scenarios, you can create a meeting without an API key — the call bills per minute to the scenario creator’s workspace credits (see Billing and credits).

curl -X POST https://api.waterr.ai/v1/meetings/end-user \
  -H "Content-Type: application/json" \
  -d '{
    "person_name": "Jane Smith",
    "scenario_id": "scenario-uuid",
    "end_user_id": "end-user-uuid"
  }'

After the meeting ends

GET /meetings/{id} returns the meeting along with its transcript, summary, and goal scores. See Post-Meeting Data.

All meeting endpoints

Method	Endpoint	Use
`POST`	`/meetings`	Create
`GET`	`/meetings/{id}`	Get one — includes transcript + analysis
`GET`	`/meetings`	List
`PUT`	`/meetings/{id}/end`	End a live meeting
`PUT`	`/meetings/{id}`	Update
`DELETE`	`/meetings/{id}`	Delete

Getting Started

Scenarios

Personas

Tools

Extend Meetings

Embed Scenario

More APIs

Guides

Integrations

Agent Skills

Meetings

Create a meeting

Override the agent for one call

Response

Joining the meeting

End-user meetings (no JWT)

After the meeting ends

All meeting endpoints

​Create a meeting

​Override the agent for one call

​Response

​Joining the meeting

​End-user meetings (no JWT)

​After the meeting ends

​All meeting endpoints

Create a meeting

Override the agent for one call

Response

Joining the meeting

End-user meetings (no JWT)

After the meeting ends

All meeting endpoints