Find answers fast

How long does it take to process a recording?

Processing runs asynchronously. Most recordings finish within a few minutes, and you get a notification or webhook when results are ready.

What languages are supported for transcription?

Transcription covers many languages, including English, Spanish, German, French, Portuguese, Mandarin, Japanese, and Arabic. Ask us for the current list for your markets. Voice biometrics works independently of language: it analyzes acoustic patterns, not words.

Do I need special hardware or software to use IdentityCall?

No. IdentityCall is entirely cloud-based. Upload recordings through our web interface, API, or integrations. For calls on virtual numbers, recordings are captured and processed automatically. No new equipment required.

What’s the minimum audio length needed for analysis?

Transcription works with any length. For voice biometrics, we need at least 3 seconds of clear speech to generate a reliable voiceprint. Longer samples improve accuracy.

What’s the accuracy rate for speaker recognition?

Match quality depends on enrollment quality, call audio, and the confidence threshold you set. Cleaner audio and longer enrollment samples improve results. Ask us for evaluation results at a stated threshold, or test on your own recordings.

How do you handle calls with multiple speakers?

Our system automatically detects and separates individual speakers within a recording. Each speaker segment is analyzed independently, allowing us to identify known speakers and flag unknown voices - even when people talk over each other.

Can voice biometrics be fooled by voice recordings or AI-generated voices?

Voice biometrics analyzes live acoustic properties that are difficult to replicate perfectly. While no system is immune to sophisticated attacks, recorded playback and current AI voice clones typically lack the natural microvariations present in live speech. We continue to improve detection of synthetic voices.

Does accent or language affect voice recognition accuracy?

No. Voice biometrics analyzes how someone speaks - the physical characteristics of their voice - not what they say or which language they use. A speaker will be recognized whether they’re speaking English, Spanish, or switching between languages.

What’s the accuracy rate for speaker recognition?

Match quality depends on enrollment quality, call audio, and the confidence threshold you set. Cleaner audio and longer enrollment samples improve results. Ask us for evaluation results at a stated threshold, or test on your own recordings.

How do you handle calls with multiple speakers?

Our system automatically detects and separates individual speakers within a recording. Each speaker segment is analyzed independently, allowing us to identify known speakers and flag unknown voices - even when people talk over each other.

Can voice biometrics be fooled by voice recordings or AI-generated voices?

Voice biometrics analyzes live acoustic properties that are difficult to replicate perfectly. While no system is immune to sophisticated attacks, recorded playback and current AI voice clones typically lack the natural microvariations present in live speech. We continue to improve detection of synthetic voices.

Does accent or language affect voice recognition accuracy?

No. Voice biometrics analyzes how someone speaks - the physical characteristics of their voice - not what they say or which language they use. A speaker will be recognized whether they’re speaking English, Spanish, or switching between languages.

What integrations are available?

We offer a native HubSpot integration and Slack notifications, with Salesforce and Pipedrive planned. Recordings can arrive via API, SFTP, or webhook from any phone system. Our REST API allows custom integrations, and webhooks notify your applications when processing completes.

How does pricing work?

Pricing is based on audio minutes processed per month. Transcription and speaker recognition are included across plans; usage limits and other capabilities vary by plan. See our pricing page for current tiers or contact sales for enterprise quotes.

Is there a free trial?

Yes. Every new account gets 5 free call analyses with no credit card required. Upload your own recordings to test accuracy on your actual call data, including voice biometrics and speaker identification.

Can I export my data?

Yes. Export transcripts, speaker profiles, and analytics in standard formats (JSON, CSV). Voice embeddings can be exported for use in your own systems. We never lock your data in - you maintain full ownership and portability.