What is Telnyx?
Telnyx is a cloud-based telecommunications platform offering programmable voice, messaging, and networking APIs. Built on a private, global IP network, Telnyx provides low-latency, carrier-grade telephony services ideal for building AI-powered voice applications, contact centers, and real-time communication solutions.
"Telnyx delivers enterprise-grade connectivity with complete API control, enabling developers to build sophisticated voice AI applications with sub-100ms latency and global reach."
Voice API
Programmable calls
SIP Trunking
Enterprise routing
Media Streaming
Real-time audio
Global Network
Private backbone
Voice API Features
Call Control API
Complete programmatic control over voice calls. Answer, transfer, bridge, record, and manipulate calls in real-time using webhooks or websockets.
Media Streaming (WebSocket)
Stream raw audio from calls via WebSocket for real-time AI processing. Bidirectional streaming enables live STT → LLM → TTS pipelines for voice AI agents.
Text-to-Speech & Speech-to-Text
Built-in TTS and STT capabilities for voice applications. Play audio, gather speech input, and transcribe calls without external services.
SIP Trunking
Enterprise-grade SIP trunking with global coverage. Connect existing PBX systems or build custom telephony infrastructure with high reliability.
Telnyx vs Twilio
Both platforms offer programmable voice APIs, but differ in architecture and pricing model.
| Feature | Telnyx | Twilio |
|---|---|---|
| Network | Private IP backbone | Public cloud |
| Latency | Lower (~50ms) | ~100ms |
| Pricing Model | Per-minute | Per-minute + fees |
| Media Streaming | ||
| SIP Trunking | ||
| Best For | Cost-sensitive, low-latency | Broad ecosystem |
Quick Start (Python)
# Install Telnyx SDK
pip install telnyx
# Make an outbound call with TTS
import telnyx
telnyx.api_key = "YOUR_API_KEY"
# Create a call
call = telnyx.Call.create(
connection_id="your-connection-id",
to="+15551234567",
from_="+15559876543",
webhook_url="https://your-server.com/webhooks/call"
)
# In your webhook handler, speak text
call.speak(
payload="Hello! Welcome to our AI assistant.",
voice="female",
language="en-US"
)
# Gather speech input
call.gather(
input_type="speech",
language="en-US",
timeout_secs=10
)
Media Streaming: For real-time AI voice agents, use
call.streaming_start()
to stream audio via WebSocket to your AI pipeline.
Use Cases
AI Voice Agents
Real-time conversational AI over phone
Contact Centers
Enterprise call routing & IVR
Video/Conferencing
WebRTC and SIP bridges
2FA / Verification
Voice OTP and verification calls
Outbound Notifications
Automated reminders & alerts
Real-Time Translation
Multilingual voice bridges