Technical guides on text-to-speech, voice AI engineering, and building production voice applications.
How Telnyx, ElevenLabs, Azure, AWS, and Google describe TTS voice metadata, where the fields disagree, and a conservative schema for normalizing them.
Andrew MunsA look inside the pronunciation API that finally gets Irish names like Niamh right in voice AI, why it happens, how the fix works, and 13 playable examples.
Andrew MunsVoice cloning uses AI to replicate a specific person's voice from a short sample. Learn how it works, how it differs from TTS, its use cases, and the consent and legal rules that govern it.
Andrew MunsNeural TTS uses deep learning to generate human-sounding speech from text. Learn how it works, why latency matters, and how infrastructure shapes voice quality.
Andrew MunsLearn how text-to-speech converts written text into natural-sounding audio, how modern neural TTS works, and why infrastructure determines voice quality.
Andrew MunsTrace the history of text-to-speech from 1769 mechanical speaking machines through neural TTS and real-time voice cloning in 2025.
Andrew MunsThe 20 best Telnyx Ultra text-to-speech voices for building voice AI agents. Persona-conditioned, expressive, and built for production. Listen to each one inline.
Andrew Muns