This site works without JavaScript.
Now in private beta · iPhone 15 Pro and later

Their voice.
Whispering
to you.

openasmr clones any voice from ten seconds — your partner's, your mother's, your own — and whispers a different ASMR session every night. Generated on your phone. Never the same twice.

· Voices stay on device · Works offline · No interruptions, ever
01 — How it works

Three quiet steps. Then a voice you love, every night.

01

Record ten seconds

A short reading, or any clip you have. Speak softly — the model picks up tone, breath, the way they say your name.

02

Pick a feeling

Sleep. Personal attention. A haircut at midnight. Twelve presets that read like a menu of small acts of care.

03

Drift

A new script every time. Whispered ambient cues — rain, scissors, breath — synced to the words. The audio renders as you listen.

02 — Twelve feelings

Stress relief, comfort, sleep, attention — the four reasons people listen.

Massage & Relaxation
body scan, slow breath
Affirmations
gentle, motivational
Sleep
slow stories, soft pauses
Personal Attention
face touching, soft words
Spa
oil, stones, hands
Haircut · Salon
scissors, comb, spray
Rain & Nature
storm, forest, ocean
Focus & Study
ambient companion
Cooking
chopping, simmer
Medical
soft questions, latex
Makeup
brushes, sprays
Tapping
wood, glass, plastic
03 — A note on consent

Your voices never leave your phone. Not for training. Not for backup. Not ever.

Voice cloning runs entirely on-device. Embeddings live in the app sandbox; the only thing the server ever sees is a request for a script — never your voice, never the audio. You confirm consent for every voice you add. Delete it, it's gone.

On-device
Whisper, voice clone, synthesis
Stateless server
Script generation only
No analytics
On voice or script content
04 — Pricing

Three tiers. Cancel any time.

Free
$0Always
Massage & Relaxation · Affirmations · Sleep
15 min sessions
1 saved voice
Live generation only
No ads, ever
Most chosen
Pro
$9.99per month
12 ASMR presets
30 min sessions
5 saved voices
Save script & voice for replay
Speed · accent · ambient mix
Offline script fallback
Max
$19.99per month
Everything in Pro
60 min sessions
Unlimited voices
Custom prompted scripts
Intimate (18+) presets
Toggle on at openasmr.ai
Intimate content stays off until you toggle it on at openasmr.ai. 18+ verification required.
05 — Listen first

Forty-five seconds. Headphones on.

A Personal Attention session in Mira's voice — one of three demo voices we've consented and licensed for the site. Yours will sound like the person you record.

Personal Attention · 0:12
Mira's voice
I'm just going to brush this piece of hair out of your eyes…
0:140:45
[breath] [hair · soft] [whisper]
06 — Quiet words

From the private beta.

"My husband travels half the month. The first night I tried it I cried, then slept eight hours. The second night I just slept."
Anya · 31 · Berlin
cloned · her partner
"I had one voicemail saved from before she passed. I wasn't sure I could handle it. The app played her like she was sitting beside me."
Jordan · 28 · Brooklyn
cloned · their late grandmother
"I cloned my own voice for affirmations. It's strange and good — like the part of me that knows better is finally the one talking."
Priya · 42 · Bangalore
cloned · herself
Featured in
The Verge
Wired
Fast Company
It's Nice That
TechCrunch
New York Mag
07 — Questions

Things people ask before they download.

Anything we missed — hello@openasmr.ai. We answer every email, slowly.

Is voice cloning legal?
Cloning a voice with consent is legal in every market we ship in. The app requires you to confirm consent for every voice you add. Cloning a public figure or someone who hasn't consented violates our terms and we'll terminate the account.
How does it work without sending my voice anywhere?
Whisper transcription, voice embedding, and TTS synthesis all run on the Apple Neural Engine. The only thing the server ever generates is the script — text — based on the preset you tap. Your voice file and the embedding never leave the phone.
Why iPhone 15 Pro and later?
On-device voice synthesis at this fidelity needs the A17 Pro Neural Engine and 8GB of unified memory. We're targeting Android and older iPhones once we can do this without quality loss.
Can I use it for someone who has passed away?
Yes — many people do. You'll need a clip you have rights to (a voicemail, a video). Be gentle with yourself; for some people it's healing, for others it surfaces grief sharply. There is no wrong answer.
What about 18+ content?
Off by default. Available on the Max tier, and only after you toggle it on at openasmr.ai with age verification. Never surfaced in the App Store build.
How are scripts generated?
A small fine-tuned language model writes them, optimized to land on the soft, second-person, present-tense register that makes ASMR work. Each script is unique; you're never hearing the same session twice.
10s
to clone a voice
On-device. Once.
unique scripts
Never the same twice.
0
data leaves your phone
Voice. Audio. Embeddings.
12
feelings to choose from
Plus a few that sleep nights.
openasmr.ai

Goodnight,
quietly.