Estimate your monthly telco carrier trunking charges, proprietary IMA SONA™ Voice Platform core fees, and TTS/STT API rates. Adjust concurrent channel loads to optimize system throughput.
Calculations assume a standard PSTN connection model. If utilizing regional numbers, flat incoming allocations and standard country prefixes apply. Outbound call rates utilize carrier lists from SIP trunking providers.
Unlike simple text bots, a real-time **Voice AI Agent** requires a sophisticated pipeline. Every turn of conversation involves four distinct backend processes occurring in sequential milliseconds:
Handles incoming and outgoing calls. Mapped using standard telecom protocols (SIP trunking) by direct carriers, charging outbound minutes.
Transcribes the user's voice into text in real-time. Deepgram's neural transcription models provide sub-50ms latency.
Processes the transcribed query and generates a response. Cost is calculated based on prompt and output token counts.
Converts the AI response text back into synthetic human speech. Optimized using ElevenLabs or Cartesia voices.
Competitors like Vapi or Bland AI add heavy markups on top of telephony and model fees. IMA’s proprietary voice infrastructure runs on raw orchestrator nodes, passing savings directly to the client:
| Feature Layer | Third-party Platforms (Vapi.ai) | IMA SONA™ Voice Core | Operational Advantage |
|---|---|---|---|
| Platform Fee / min | $0.150 / minute (Flat) | $0.060 / minute (Volume scaled) | 60% Infrastructure saving |
| Data Residency | Shared US servers (No HIPAA control) | On-premise / Private AWS Cloud | Full HIPAA & GDPR compliance |
| SIP Trunk support | Standard Twilio/Plivo binds | Direct Carrier Tier-1 peering | Zero carrier markups |
| Cost Layer | Global Standard (USD) | Indic / Local (INR) | Primary Provider |
|---|---|---|---|
| Platform Middleware | $0.060 / minute | ₹5.0 / minute | IMA SONA™ Core |
| Outbound Telephony | $0.013 / minute | ₹1.00 / minute | Telco Carrier |
| Speech-to-Text (STT) | $0.012 / minute | ₹0.10 / minute | Deepgram / Sarvam AI |
| Text-to-Speech (TTS) | $0.015 / minute | ₹0.10 / minute | Cartesia / Sarvam AI |
Latency is the delay between a user finishing speaking and the AI responding. We minimize this by using streaming WebSockets and WebSocket APIs directly connected to GPU hosting nodes.
Yes. IMA specializes in connecting custom SIP trunks and virtual numbers from carriers like Tata Communications, Airtel, and international telco routes.
By selecting regional options, we deploy multilingual STT models (such as Sarvam's Shruthi model) that handle code-switching and Hinglish seamlessly.
The system only charges for the minutes the call remains connected to our voice gateway during the human transfer bypass. External transfer rates depend on your trunk carrier.