Question 1

How do I run Sayna without API keys for local testing?

Accepted Answer

Use audio-disabled mode by sending a WebSocket configuration message with 'audio: false'. This allows testing control flows and WebSocket messaging without setting up provider credentials, as described in the Quick Start section.

Question 2

Sayna vs using direct provider APIs: what's the advantage?

Accepted Answer

Sayna provides a unified interface across multiple providers, real-time streaming, and integrated features like noise filtering and VAD. Direct APIs are simpler for single-provider use but lack Sayna's abstraction and advanced processing capabilities.

Question 3

Can Sayna handle group voice calls or conferences?

Accepted Answer

Yes, through LiveKit integration for WebRTC audio streaming with room-based communication. This enables multi-party scenarios, and SIP configuration can extend it to phone call systems with proper setup.

Question 4

How to enable noise filtering in Sayna?

Accepted Answer

Enable the 'noise-filter' feature flag during build or use the default Docker image. Configure audio processing pipelines in WebSocket messages to apply DeepFilterNet noise suppression for cleaner audio input.

Question 5

What authentication methods does Sayna support for production?

Accepted Answer

Supports JWT-based external authentication services, API key headers, or query parameters. Setup involves environment variables like AUTH_SERVICE_URL and signing keys, with detailed steps in the Authentication section.

Question 6

Is Sayna suitable for low-latency real-time applications?

Accepted Answer

Yes, with WebSocket-based bidirectional streaming and optimized audio buffering, it's designed for low latency. However, performance depends on server resources and optional CPU-intensive features like DeepFilterNet.

Sayna

What is Sayna?

Overview

Use Cases

Best For

Related Projects

Found a gem we're missing?

Not Ideal For

Pros & Cons

Pros

Cons

Frequently Asked Questions