Skip to main content
POST
/
v1
/
sessions
/
{id}
/
analysis
curl -X POST https://api.kallglot.com/v1/sessions/sess_01HXYZ123456789/analysis \
  -H "Authorization: Bearer sk_live_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "analyses": ["sentiment", "summary", "action_items", "quality"],
    "language": "en",
    "webhook_url": "https://your-app.com/webhooks/analysis"
  }'

{
  "id": "ana_01HDEF456789012",
  "object": "analysis",
  "session_id": "sess_01HXYZ123456789",
  "status": "queued",
  "analyses": ["sentiment", "summary", "action_items", "quality"],
  "language": "en",
  "created_at": "2026-03-26T11:10:00Z",
  "estimated_completion": "2026-03-26T11:10:30Z"
}
Request AI-powered analysis of a completed session. Analysis includes sentiment detection, topic extraction, action items, and quality scoring.

​
Path Parameters

​
id
string
required
The unique session identifier (e.g., sess_01HXYZ123456789).

​
Request Body

​
analyses
array
required
Types of analysis to perform. Include one or more of:
  • sentiment - Overall and per-segment sentiment analysis
  • topics - Topic and keyword extraction
  • summary - AI-generated conversation summary
  • action_items - Extract action items and follow-ups
  • quality - Call quality scoring (talk time, interruptions, etc.)
  • compliance - Compliance and script adherence checking
​
language
string
Language for the analysis output. Defaults to en. Supports any language the session was conducted in.
​
webhook_url
string
URL to receive the analysis results when complete. If not provided, poll the analysis endpoint.
​
custom_prompts
object
Custom prompts for specific analysis types.

​
Response

​
id
string
Unique analysis identifier (e.g., ana_01HDEF456789012).
​
object
string
Always analysis.
​
session_id
string
The session being analyzed.
​
status
string
Analysis status:
  • queued - Analysis is queued for processing
  • processing - Analysis is in progress
  • complete - Analysis is complete
  • failed - Analysis failed
​
analyses
array
Requested analysis types.
​
created_at
string
Request timestamp (ISO 8601).
​
estimated_completion
string
Estimated completion time (ISO 8601).
curl -X POST https://api.kallglot.com/v1/sessions/sess_01HXYZ123456789/analysis \
  -H "Authorization: Bearer sk_live_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "analyses": ["sentiment", "summary", "action_items", "quality"],
    "language": "en",
    "webhook_url": "https://your-app.com/webhooks/analysis"
  }'

{
  "id": "ana_01HDEF456789012",
  "object": "analysis",
  "session_id": "sess_01HXYZ123456789",
  "status": "queued",
  "analyses": ["sentiment", "summary", "action_items", "quality"],
  "language": "en",
  "created_at": "2026-03-26T11:10:00Z",
  "estimated_completion": "2026-03-26T11:10:30Z"
}

​
Analysis Types

​
Sentiment Analysis

Detects the emotional tone of the conversation:
  • Overall sentiment score (-1 to 1)
  • Per-segment sentiment
  • Sentiment trend over time
  • Key moments (positive/negative peaks)

​
Summary

AI-generated summary of the conversation:
  • Brief overview (1-2 sentences)
  • Key discussion points
  • Outcome/resolution

​
Action Items

Extracts commitments and follow-ups:
  • Action items mentioned
  • Assigned party (agent or customer)
  • Due dates if mentioned
  • Priority level

​
Quality Scoring

Call quality metrics:
  • Talk time ratio (agent vs customer)
  • Interruption count
  • Hold time
  • Response latency
  • Overall quality score

​
Compliance

Script adherence and compliance checking:
  • Required phrases used
  • Prohibited phrases detected
  • Disclosure completion
  • Overall compliance score

​
Notes

Analysis is performed asynchronously. Use webhooks for notification or poll the analysis endpoint.
Analysis typically completes within 30 seconds. Complex analyses or long sessions may take up to 2 minutes.
Analysis requires a completed session with transcription enabled. Sessions without transcripts cannot be analyzed.