Skip to main content

​
Overview

The Telepath dashboard provides real-time monitoring and historical analysis of your Voice AI calls. Access the dashboard at: https://dashboard.telepathvoice.com

​
Dashboard Features

​
Connections

Manage all your Voice AI connections in one place. View Connections:
  • Connection name and status
  • Associated AI provider (OpenAI, ElevenLabs, Custom)
  • SIP username and authentication
  • Recent call count and success rate
Actions:
  • Edit agent configuration
  • Rotate SIP password
  • View recent calls
  • Delete connection
Create separate connections for different AI agents, allowing you to A/B test different approaches or maintain separate integrations.

​
Call Log

View all inbound and outbound calls with detailed metrics. Call Details Include:
  • Call ID: Unique identifier for tracking
  • Timestamp: When the call connected
  • Duration: Total call length
  • Status: Connected, Disconnected, Failed, etc.
  • Carrier Lag: Delay from carrier to Telepath
  • AI Latency: Time for AI agent to respond
  • Gateway Processing: Telepath’s bridge time
Filtering:
  • By connection
  • By date range (up to 30 days)
  • By status (connected, failed, etc.)
  • By duration
Sorting:
  • Most recent first
  • Longest duration
  • Highest latency
  • Highest carrier lag

​
Call Details (Per-Call Inspection)

Click any call in the log to view comprehensive details. Overview:
  • Call ID and timestamp
  • Total duration and status
  • Connected to which connection/AI agent
Metrics:
  • Carrier Lag: Delay from carrier to Telepath
  • AI Latency: Average AI response time
  • Gateway Processing: Telepath processing time
  • Packet Loss: Percentage of dropped packets
  • Jitter: Audio timing variance
Turn-by-Turn Analysis:
  • Caller spoke for X seconds
  • AI processed for Y milliseconds
  • AI responded for Z seconds
  • Total turn duration
Carrier Lag high? Your carrier might be overloaded.AI Latency high? Your AI provider might be slow.Gateway Processing high? Contact support.
SIP Trace:
  • Complete SIP signaling log
  • Useful for diagnosing connection issues
  • Shows all SIP INVITE, 200 OK, BYE, etc.
Audio Quality:
  • Codec used (G.711, G.722)
  • Sample rate (8kHz, 16kHz)
  • Jitter buffer events
  • Silence detection

​
Analytics

View aggregated performance metrics across all calls. Overview Stats:
  • Total calls (today/week/month)
  • Total minutes (today/week/month)
  • Average call duration
  • Success rate (% completed calls)
  • Average latency
Trends:
  • Calls per hour/day/week
  • Average duration trends
  • Latency trends
  • Error rate trends
Provider Comparison:
  • Metrics by AI provider
  • Performance of each integration
  • Success rates by connection
Geographic Insights (if available):
  • Calls by region
  • Latency by location
  • Carrier performance by region

​
Settings

Manage account and notification preferences. Account Settings:
  • Email address
  • Time zone
  • Default connection
  • Notification preferences
API Keys:
  • Generate and revoke API keys
  • View key usage
  • Set key permissions
Billing (if applicable):
  • Current usage
  • Billing period
  • Payment method
  • Usage alerts
Data Retention:
  • Current retention period (30 days default)
  • Data export options

​
Key Metrics Explained

​
Carrier Lag

Time from when the carrier initiates a call to when Telepath receives the audio stream. Typical Range: 50-200ms What’s Normal: 50-100ms is ideal High Values: May indicate carrier congestion

​
AI Latency

Time for your AI agent to process audio and generate a response. Typical Range:
  • OpenAI: 100-300ms
  • ElevenLabs: 200-500ms
  • Custom: Depends on implementation
What’s Normal: Varies by provider High Values: May need to optimize AI agent

​
Gateway Processing

Telepath’s internal processing time (audio conversion, routing, etc.). Typical Range: 10-50ms What’s Normal: <20ms is excellent High Values: Rare, contact support if consistently high

​
End-to-End Latency

Total time from caller speech to AI response: 
Carrier Lag + AI Latency + Gateway Processing
Target: <500ms for natural conversation

​
Monitoring Best Practices

​
Daily

  • Check for any failed calls
  • Verify success rate >95%
  • Monitor for any anomalies

​
Weekly

  • Review latency trends
  • Check for codec mismatches
  • Analyze performance by provider
  • Review error logs

​
Monthly

  • Analyze usage patterns
  • Review carrier performance
  • Optimize AI agent if needed
  • Plan for future capacity

​
Setting Alerts

Enable notifications for:
  • Failed calls (more than 5%)
  • High latency (>500ms average)
  • Low success rate (<95%)
  • Carrier connection issues

​
Troubleshooting via Dashboard

​
Call Dropped Immediately

  1. Click the call in the log
  2. View the SIP trace
  3. Look for disconnect reason
  4. Check Carrier Lag (excessive delay?)
  5. Check AI agent status

​
Poor Audio Quality

  1. View call details
  2. Check codec used
  3. Review jitter buffer events
  4. Check packet loss %
  5. Consider switching to G.722

​
High Latency

  1. Identify which component is slow:
    • High Carrier Lag? → Carrier issue
    • High AI Latency? → AI provider issue
    • High Gateway Processing? → Contact support
  2. Review trends over time
  3. Check if pattern specific to one provider

​
Silent Calls

  1. View call details
  2. Check if AI agent was reached
  3. Review SIP trace for errors
  4. Verify AI provider credentials
  5. Test with a different connection

​
Exporting Data

​
View Call History

  1. Go to Call Log
  2. Select date range
  3. Apply filters
  4. Click Export (if available)
  5. Choose format (CSV, JSON)

​
Accessing Raw Data via API

See API Reference for programmatic access to call data.

​
Tips & Tricks

Custom Dashboard URL:
  • Bookmark specific date ranges
  • Share filtered views with team
Real-time Monitoring:
  • Keep dashboard open during testing
  • Set browser to auto-refresh
  • Use full-screen mode
Comparison Analysis:
  • Filter to specific connections
  • Compare performance over time
  • Identify best-performing agents
Capacity Planning:
  • Monitor call volumes
  • Track growth trends
  • Plan for scaling

​
API Reference

Monitor your system programmatically:

List Calls

Retrieve call history programmatically

Get Call Details

Access detailed metrics for specific calls

Analytics Overview

Query aggregated analytics