MurmurAI closes $20M to scale adaptive voice models for games

AI · 5 min read

MurmurAI's adaptive voice models condition vocal performance on emotion, pacing, and narrative context to produce varied NPC lines that match in-game events. The SDK allows runtime control signals so audio directors can set intensity, sarcasm, or tiredness and combine generated lines with pre-recorded assets for hybrid pipelines.

A $20 million funding round led by Harbor Lane will help MurmurAI expand language support, improve latency, and produce tools for voice directors to tweak model behavior. Early-access partners include several mid-sized RPG studios experimenting with live event voices and reactive companion characters.

Audio leads noted the potential to generate contextual variants for thousands of lines without hiring large voice casts, but stressed the importance of quality gates and human review to maintain character consistency. MurmurAI plans to add workflow integrations for auditioning and approving generated performances.