Skip to Content
Frequently Asked Questions

Frequently Asked Questions

Common questions about IdentityCall’s API, features, and implementation.

General Questions

What is IdentityCall?

IdentityCall is a call recording analysis platform that provides:

  • Transcription: Convert audio to text with speaker diarization
  • Emotion Detection: Analyze emotional tone per speaker
  • Goal Evaluation: AI-powered assessment of call objectives
  • Pause Analysis: Detect significant pauses for compliance
  • Voice Biometrics: Speaker identification via voice profiles

What audio formats are supported?

FormatExtensionMax Size
MP3.mp3100 MB
WAV.wav500 MB
M4A.m4a100 MB
OGG.ogg100 MB
FLAC.flac500 MB
WebM.webm100 MB

What languages are supported?

IdentityCall uses Google Gemini for transcription, supporting 100+ languages. Common languages include English, Spanish, French, German, Italian, Portuguese, Chinese, Japanese, Korean, Arabic, Hindi, and Russian.

See the complete language list.

How long does transcription take?

Processing time depends on audio length:

  • Under 5 minutes: 30-60 seconds
  • 5-15 minutes: 1-3 minutes
  • 15-60 minutes: 3-10 minutes
  • Over 60 minutes: 10-20 minutes

Transcription runs asynchronously. Upload your file, receive an ID, and poll for completion or use webhooks.

Authentication & Access

How do I get an API key?

API keys are generated from your IdentityCall dashboard:

  1. Log in at app.identitycall.ai 
  2. Navigate to Settings > API Keys
  3. Click “Create API Key”
  4. Set permissions (read, write, delete)
  5. Copy and securely store your key

Keys are prefixed with idc_ and should be passed as Bearer tokens.

What permissions do API keys have?

API keys support three permission levels:

  • read: List and view recordings, transcriptions, results
  • write: Upload new recordings, update metadata
  • delete: Remove recordings permanently

Set minimum required permissions for security.

How do I authenticate requests?

Include your API key in the Authorization header:

Authorization: Bearer idc_your_api_key

What happens if my API key is compromised?

  1. Immediately revoke the key in your dashboard
  2. Create a new key with appropriate permissions
  3. Update your applications with the new key
  4. Review API logs for unauthorized access

Transcription & Analysis

How accurate is the transcription?

Transcription accuracy depends on audio quality:

  • Clear audio: 95%+ accuracy
  • Background noise: 85-95% accuracy
  • Multiple speakers talking: 80-90% accuracy

For best results, use recordings with clear audio and minimal background noise.

How does speaker diarization work?

Speaker diarization identifies and labels different speakers in a recording. IdentityCall:

  1. Detects voice changes throughout the audio
  2. Clusters similar voice segments
  3. Assigns consistent labels (Speaker 1, Speaker 2, etc.)
  4. Optionally matches against enrolled voice profiles

What emotions are detected?

Each dialogue segment includes probability scores for:

  • Happy
  • Neutral
  • Calm
  • Sad
  • Angry
  • Fearful
  • Surprised
  • Disgust

Scores range from 0.0 to 1.0 and sum to approximately 1.0.

How do goals work?

Goals are configured in your project settings with:

  • Name: Descriptive title
  • Description: What constitutes achievement
  • Criteria: AI evaluation rules

After transcription, the AI evaluates each goal and returns:

  • Whether the goal was met (boolean)
  • Confidence score (0-100)
  • Explanation of the evaluation

What makes a pause “non-compliant”?

Pause compliance is determined by your project’s configured thresholds:

  • Minimum duration: Pauses shorter than this are ignored
  • Maximum duration: Pauses longer than this are flagged
  • Context rules: Specific requirements for regulated scenarios

Non-compliant pauses may indicate missed disclosures or inappropriate wait times.

Voice Biometrics

How do voice profiles work?

Voice profiles capture unique voice characteristics:

  1. Enrollment: Upload audio of a known speaker
  2. Embedding: Extract voice signature
  3. Matching: Compare against future recordings

When processing new recordings, IdentityCall matches speaker segments against enrolled profiles.

How many samples are needed for enrollment?

Minimum requirements:

  • Duration: At least 10 seconds of clear speech
  • Quality: Clear audio without background noise
  • Variety: Natural speaking (not reading scripted text)

More samples improve matching accuracy.

How accurate is voice matching?

Voice biometric accuracy varies:

  • Same conditions: 95%+ accuracy
  • Different phones/microphones: 85-95% accuracy
  • Background noise: 75-90% accuracy

Matching returns confidence scores to help set thresholds.

API Usage

What are the rate limits?

Rate limits depend on your subscription:

  • Starter: 60 requests/minute
  • Professional: 300 requests/minute
  • Enterprise: Custom limits

Rate limit headers are included in all responses:

X-RateLimit-Limit: 60
X-RateLimit-Remaining: 45
X-RateLimit-Reset: 1642345678

How do I handle rate limiting?

When rate limited (HTTP 429), the response includes:

  • retry_after: Seconds to wait before retrying

Implement exponential backoff:

import time
 
def make_request_with_retry(func, max_retries=3):
    for attempt in range(max_retries):
        try:
            return func()
        except RateLimitError as e:
            if attempt < max_retries - 1:
                time.sleep(e.retry_after)
            else:
                raise

Is there a webhook for completion notifications?

Yes, webhooks can notify your application when processing completes. Configure webhook URLs in your project settings. Webhook payloads include:

  • Recording ID
  • Status (completed, failed)
  • Timestamp

How do I handle pagination?

List endpoints support pagination:

GET /recordings?page=1&per_page=50

Responses include metadata:

{
  "meta": {
    "current_page": 1,
    "total_pages": 10,
    "total_count": 487,
    "per_page": 50
  }
}

Iterate through pages until current_page >= total_pages.

Can I delete recordings after processing?

Yes, use the DELETE endpoint:

DELETE /recordings/{id}

This permanently removes:

  • Audio file
  • Transcription
  • Analysis results
  • Associated metadata

This action cannot be undone.

Security & Privacy

How is my data secured?

IdentityCall implements multiple security measures:

  • Encryption in transit: TLS 1.3 for all API connections
  • Encryption at rest: AES-256 for stored data
  • Access controls: Role-based permissions
  • Audit logging: All API access is logged

Where is data stored?

Data is stored in secure cloud infrastructure:

  • Primary: EU (Germany)
  • Backups: Geographically distributed

Enterprise plans support data residency requirements.

How long is data retained?

Default retention periods:

  • Audio files: 90 days
  • Transcriptions: Until deleted
  • Analysis results: Until recording deleted

Enterprise plans support custom retention policies.

Is IdentityCall GDPR compliant?

Yes, IdentityCall is designed for GDPR compliance:

  • Data processing agreements available
  • Right to deletion supported
  • Data export functionality
  • Audit logs maintained

Contact support for compliance documentation.

Troubleshooting

Why did my transcription fail?

Common failure reasons:

  • Invalid audio format: Use supported formats
  • Corrupted file: Re-encode the audio
  • Too short: Minimum 1 second of speech
  • No speech detected: Check audio contains speech

Check the recording’s error details in the response.

Why are speakers not identified?

Speaker identification requires:

  • Enrolled profiles: Create voice profiles first
  • Sufficient audio: At least 5 seconds per speaker
  • Clear audio: Background noise affects matching

Why are emotions showing unexpected values?

Emotion detection may be affected by:

  • Audio quality: Background noise skews results
  • Speaker volume: Very quiet speech is harder to analyze
  • Non-speech sounds: Laughter, coughing affect scores

How do I debug API requests?

  1. Check response status codes and error messages
  2. Verify authentication header format
  3. Confirm request body/parameters are valid
  4. Review rate limit headers
  5. Contact support with request IDs for assistance

Integration

Do you have SDKs?

Yes, SDKs are available for:

SDKs handle authentication, retries, and error handling.

Can I use the API with my existing CRM?

Yes, the REST API integrates with any system that can make HTTP requests. Common integrations:

  • Salesforce: Via Apex HTTP callouts
  • HubSpot: Via workflow webhooks
  • Zendesk: Via custom apps

Is there a Zapier integration?

Not currently. Use the REST API directly or webhooks for automation.

Can I process recordings in real-time?

The current API processes completed recordings. For live call analysis, contact sales about our real-time streaming solution.

Still Have Questions?

Email our support team

Contact Support

Complete endpoint documentation

API Reference
Ruby