Skip to content
AImpact
IT EN
High Voice & Audio · 1 min read

Sesame Maya & Miles: AI voices that 'think aloud' cross the uncanny valley

In one sentence Sesame (founded by former Oculus/Meta engineers) ships Maya and Miles, conversational voices with prosody, hesitations, and breaths so natural they trigger the 'feels like a real person' effect. Base CSM-1B model open Apache 2.0.

Needs review Official source
ShareLinkedInX
Reading level

Sesame, a startup founded by ex-Oculus engineers, ships Maya and Miles: two AI voices for conversations that sound surprisingly real. Not only do they speak fluently, they breathe, hesitate, chuckle, say "uhm" and self-correct — the small human imperfections AI voices usually scrub.

Free to try on sesame.com, the effect goes viral fast: thousands of X/Reddit users say "I had the weirdest conversation, I forgot it was AI". Sesame names the problem the "uncanny valley of voice" — and claims to have crossed it.

In parallel they release CSM-1B, the base model in open source (Apache 2.0), so the community can experiment. Opens the new generation of voice AI where prosody quality matters more than single MOS benchmark numbers.

Companies

Sesame

Tools

Sesame Maya, Sesame Miles, CSM-1B

Tags

SesameConversational VoiceCSMMayaMilesOpen Source

Sources