Hi Joaquim pedro de, When building with ElevenLabs, choosing the right API type can help you optimize for quality, latency, or flexibility — depending on your use case. Here's a quick breakdown: Standard API Designed for high-quality, non-real-time workloads like audiobook creation or article readovers. Supports full audio generation via TTS, STS, Voice Isolator, and Voice Generation. Output is returned as a single audio file. Streaming API Built for near-real-time playback. Returns audio in chunks as it's generated — ideal when you have the full input upfront but want faster playback. Compatible with our TTS, Voice Changer, and Audio Isolation products. Node and Python SDKs simplify stream handling. Read more here WebSockets Purpose-built for real-time conversations. Sends and receives audio in a continuous stream, enabling truly interactive AI voice agents. Best used in applications where latency is critical — like conversational AI systems. Read more here Each option offers different performance characteristics and is tailored to specific workflows. If you're unsure which fits best, we're happy to help you decide.
| 
