Pingr AI Documentation
Pingr AI gives engineering and product teams ultra-low latency conversational voice AI infrastructure. Connect SIP trunks, WebRTC bidirectional audio streams, intelligent conversation workflows, and multi-channel automations (WhatsApp & SMS) with zero operational friction.
Pingr AI Architecture — Voice AI, WhatsApp Business API, and Real-Time Telephony
Quick Navigation
Quickstart
Create your first voice agent and place a live test call in under 5 minutes.
Authentication
Generate and manage scoped Bearer tokens to secure your API requests.
Voice Agents
Configure system prompts, LLM parameters, neural voices, and tools.
API Reference
Explore the full REST API specification, pagination, and status codes.
Developer Console & API Access
Manage voice agents, view live call transcripts, configure webhooks, and generate scoped credentials directly from the Pingr AI Console.
Pingr AI Developer Portal — API Key and Workspace Management
Provisioning Credentials
- Create Scoped Keys: Generate dedicated keys with granular permissions (
agents:read,calls:dispatch,campaigns:manage). - Set Expiration: Protect mission-critical deployments with automatic key rotation or non-expiring production secrets.
Creating a new scoped API credential in the Pingr AI Console
- Secure Your Secret: Copy and store your secret key immediately in your environment variables or secret manager.
API key generated — copy and securely store in your environment
Architecture Highlights
- Sub-500ms Audio Latency: Streaming speech-to-text (Deepgram Nova-2), LLM token streaming (Claude 3.5 Sonnet / GPT-4o), and neural text-to-speech pipelines (Cartesia & ElevenLabs).
- Bi-directional Interruption Handling: Real-time conversational barge-in allows users to interrupt natural speech smoothly.
- Multi-channel Execution: Automatically dispatch WhatsApp follow-ups, trigger webhooks, and log structured transcripts into your CRM.