Post-Meeting Data

Once a meeting ends, everything you need — summary, transcript, goal scores, strengths, growth areas, speech analytics — is returned by a single endpoint:

GET /meetings/{meetingId}

You only need a second call if you want the recording media URL.

The single call

curl https://api.waterr.ai/v1/meetings/{meetingId} \
  -H "Authorization: Bearer wai_<your_key>"

The response includes:

Field	What it is
`status`	`created` / `active` / `ended`
`analysis_status`	`not_started` / `processing` / `completed` — poll this
`summary`, `ended_at`, `duration_seconds`	Meeting-level metadata
`Scenario`	The scenario this meeting ran (with nested `Persona` and `Goals`)
`sentenced_transcripts[]`	Full transcript — `speaker`, `message`, `start_duration`, `sequence_number`
`analysis`	The full AI analysis block (only present once `analysis_status: "completed"`)
`participants[]`	Who joined the room

The analysis block contains:

summary, call_summary — short and long-form summaries
average_score, total_score — overall scoring
strengths, growth_areas — qualitative feedback
filler_words, weak_words, repetitions, sentence_starters — speech analytics
goal_results[] — per-goal score with short_description, long_description, score, and the original goal definition

Polling for analysis

Analysis is generated 30–60 seconds after the meeting ends. Poll GET /meetings/{id} and check analysis_status:

async function waitForAnalysis(meetingId, maxAttempts = 12) {
  for (let i = 0; i < maxAttempts; i++) {
    const res = await fetch(`https://api.waterr.ai/v1/meetings/${meetingId}`, {
      headers: { 'Authorization': `Bearer ${API_KEY}` }
    });
    const meeting = await res.json();

    if (meeting.analysis_status === 'completed') return meeting;
    if (meeting.analysis_status === 'not_started' && meeting.status !== 'ended') {
      throw new Error('Meeting has not ended yet');
    }

    await new Promise(r => setTimeout(r, 5000));
  }
  throw new Error('Analysis not ready after max attempts');
}

const meeting = await waitForAnalysis('meeting-uuid');
console.log('Score:', meeting.analysis.average_score);
console.log('Transcript:', meeting.sentenced_transcripts);
console.log('Goal results:', meeting.analysis.goal_results);

For production, use Webhooks (meeting.ended, analysis.completed) instead of polling.

Recording (separate call)

The transcript and analysis are inline. The recording media URL lives on the CDN and is fetched separately:

curl https://api.waterr.ai/v1/recordings/url-with-thumbnail/{meetingId} \
  -H "Authorization: Bearer wai_<your_key>"

The recording is typically ready 2–5 minutes after the meeting ends — slower than analysis. You’ll get back a CloudFront URL plus a thumbnail. If you just want the raw URL without a thumbnail, use GET /recordings/meeting/{meetingId} instead.

Re-running analysis

If a meeting ended but analysis failed (or you’ve updated the scenario’s goals and want to re-score), trigger a fresh run:

curl -X POST https://api.waterr.ai/v1/analyses/trigger/{meetingId} \
  -H "Authorization: Bearer wai_<your_key>"

Underlying endpoints

If you’d rather hit the individual resources directly:

Data	Endpoint
Analysis only	`GET /analyses/meeting/{meetingId}`
Transcript only	`GET /transcripts/meeting/{meetingId}`
Recording URL	`GET /recordings/url-with-thumbnail/{meetingId}`
Re-trigger analysis	`POST /analyses/trigger/{meetingId}`

But for almost every integration, GET /meetings/{meetingId} is what you want.

Next steps

Webhooks

Get notified the moment analysis completes — skip polling.

Session Lifecycle

Understand the full flow from create → join → end → analysis.

Getting Started

Meetings

Personas

Scenarios

Embed Scenario

Goals

Avatars

Voices

Tools

Users

Guides

Integrations

Agent Skills

Post-Meeting Data

The single call

Polling for analysis

Recording (separate call)

Re-running analysis

Underlying endpoints

Next steps

Webhooks

Session Lifecycle

​The single call

​Polling for analysis

​Recording (separate call)

​Re-running analysis

​Underlying endpoints

​Next steps

Webhooks

Session Lifecycle

The single call

Polling for analysis

Recording (separate call)

Re-running analysis

Underlying endpoints

Next steps