Find answers fast
Everything you need to know about IdentityCall.
FAQ
Step-by-Step feature guides
Recording & analyzing a call
1. Upload your audio file via the dashboard or receive an inbound call on your virtual number. 2. The system automatically processes the audio through our AI pipeline (ASR -> Diarization). 3. Within minutes, view the transcription, speaker breakdown, and emotional analysis in the 'Recordings' tab.
Setting up a voice agent
1. Go to 'Voice Agents' and click 'Create Agent'. 2. Define your agent's persona and instructions in the prompt editor. 3. Assign a virtual number to the agent. 4. Call the number to test the agent's responses and latency in real-time.
Understanding voice ID principles
1. Enrollment: Upload a clear sample of a speaker's voice to create a 'Voice Profile'. 2. Comparison: When a new call comes in, the system compares the audio against all active profiles. 3. Match: If the voice matches a profile with high confidence, the speaker is automatically labeled in the transcript.
Technical troubleshooting
How to debug webhook signature failures?
Ensure you are validating the 'Telnyx-Signature-Ed25519' header using the public key available in your account settings. Reject any requests where the signature verification fails.
Why is my voice profile returning low confidence scores?
Low confidence often results from poor audio quality (background noise, cross-talk) or short sample duration. Ensure enrollment samples are at least 30 seconds of clear, isolated speech.
What is the pause detection threshold?
By default, the system flags silence longer than 3 seconds as a 'pause'. You can adjust this threshold in your Project Settings under 'Analytics Configuration'.
Still have questions?
Start your free 14‑day trial - no credit card required.
