AI ASMR voices: close-mic texture without a microphone

ASMR is a proximity medium. The whole genre is the illusion that someone is very close to your ear, which is why the standard AI voice, engineered for broadcast clarity, lands wrong: too loud, too clean, no air.

Close-mic texture is the difference. It's the sound of breath near the capsule, the soft attack of consonants, the tiny room tone that tells the ear "this is intimate, this is near."

▶ Listen to Whisper

What the whisper voice does

The whisper voice is built for exactly this: intimate, close-mic, and softly breathy. It doesn't flatten quiet speech into a quiet monotone. It keeps the air, the sibilance, and the pacing that make whispered lines work.

For longer ASMR content

• Angel: soft, delicate cadence with gentle warmth. The comfort voice.
• Honey: sweet and comforting, irresistibly warm. The unwind-at-midnight voice.
• Chill: laid-back and smooth. The slow, even sleep narration.

▶ Listen to Angel

Using it in an app

It's the same API as everything else: POST /v1/text-to-speech/whisper returns MP3, PCM or Opus, and the realtime WebSocket streams frame-by-frame for live interactive sessions. Generate a whole library of close-mic lines at 1 credit per character. At 800 characters per minute, an hour of ASMR runs about 48,000 credits.

▶ Listen to Honey

The microphone stays optional. The proximity is the product.