1. Use ElevenLabs AI text-to-speech to create voice for your game characters right in Unreal Editor. Import Data Table with list of phrases from CSV or JSON (or create it in the UE) and convert it to set of SoundWave asset by two clicks. Then, re-generate assets separately to get desired intonation.
2. Generate speech from text in runtime, just like with any other popular TTS service.

Video - YouTube

Runtime generation supports streaming in two modes: via HTTP (to provide full input text and receiva audio in chunks) and via web socket (to provide input text by chunks).