01 · AWKWARD SILENCE
Four seconds in. They’ll never know you blanked.
The room goes quiet. Hearby waits the awkward beat, then whispers the line that fits where the conversation was going. You sound like you were always thinking.
UNIVERSAL AMBIENT AI · IN YOUR EAR
Smartest in the room.
Instantly.
Tap Listen. Pop your AirPods in. Hearby reads the room, pulls the smartest line from the web, your calendar, and everyone you’ve ever met — and whispers it before they finish the sentence.
Listening for an awkward silence
BUILT ON
- AirPods
- Whisper
- Claude · Haiku 4.5
- Cloudflare
- ElevenLabs
- Stripe
AS SEEN ON
- TikTok
- Product Hunt
- X
- Hacker News
- The Information
FIVE EDGES · TUNED ON REAL CONVERSATIONS
Every moment has
a line. Hearby finds it.
Five triggers. Each one a moment Hearby listens for — the silence, the name dropped, the question landed, the memory worth recalling, the room shifting. You stay present. Hearby delivers the line.
02 · NAME DROP
Drop a name. Own the room.
“Sarah from Stripe” lands and you already have her role, her last big launch, and the one fact she wants you to know. Reference it. Watch her face change.
03 · QUESTION RESCUE
Caught without an answer? Not anymore.
Any question, instantly drafted. Use it word-for-word, twist it your way, or ignore it. Never again caught with “uhh”.
04 · MEMORY RECALL
Remember everything they ever told you.
Every conversation searchable for 24 hours. The dog’s name, the deal she’s chasing, the city he just moved to. Hearby resurfaces it the second it matters.
05 · TONE READ
Read the room. Win it.
Voice stress, pause patterns, micro-shifts in tone. Hearby reads them all and tells you when to push, when to soften, when to shut up.
≤7 WORDS · HARD CAPPED
Seven words.
Surgical.
The cap isn’t a limit — it’s the weapon. Anything longer and you stop listening to the person in front of you. Seven words land. They never see it coming.
01 · HEARD
“so I’ve been doing the vet thing for about ten years now”
02 · DECIDED
03 · WHISPERED
ask how she got into vet work
On-device speech recognition. Backend never sees a quoted word.
PRIVACY BY ARCHITECTURE
What others say
stays on your device.
Full stop.
Every privacy boundary is enforced in the architecture, not in the marketing. The other person’s audio never leaves your phone. Transcripts live in RAM and are dropped each chunk. The backend only ever sees extracted facts and your own calendar context — never quoted speech.
On-device
Backend
At rest
Other-party audio
✓
—
—
Other-party transcript
✓
—
—
Extracted facts
✓
✓
—
Your calendar event title
✓
title
—
Whisper text returned
—
stream
—
Audit metadata
—
✓
90 days
Two-party-consent states detected automatically — CA · FL · IL · MD · MA · MT · NV · NH · PA · WA. Disclosure Mode announces processing before the first word; Solo Mode discards the other voice pre-STT.
Read the full architecturePRICING · ONE TIER
Twenty a month.
Smartest move you’ll make.
Same product, same whispers, same privacy. Pick the cadence. Become the room.
MONTHLY
$20/month
- 14-day free trial
- iOS via App Store
- Cancel anytime
- Same product, same whispers
FOUNDER RATE · SAVE 59%
ANNUAL
$99/year
- 14-day free trial
- Web checkout via Stripe
- Billed once · cancel anytime
- Same product, same whispers
UNLIMITED WHISPERS · ON-DEVICE STT · 24-HOUR CONVERSATION MEMORY · DISCLOSURE & SOLO MODES · CANCEL ANYTIME
QUESTIONS, HONESTLY ANSWERED
The stuff people actually ask.
Is this legal where I live?
In one-party-consent states, you can record conversations you’re part of. In the ten two-party-consent states (CA, FL, IL, MD, MA, MT, NV, NH, PA, WA), Hearby auto-detects your location and switches to Disclosure Mode (announces processing audibly before the first word) or Solo Mode (drops the other voice before any STT runs). You stay legal everywhere we ship.
Does it record other people?
No. Other-party audio and transcripts live in RAM on your iPhone and are dropped each chunk. Nothing the other person says ever leaves your device. The backend only receives extracted facts — a role, an employer, a city — and your own calendar context. Never quoted speech.
Will the other person know?
In one-party states, no — you don’t owe a disclosure. In two-party states, Disclosure Mode plays a short audible announcement before the first user utterance at ~30dB above ambient. Solo Mode doesn’t announce — it just refuses to transcribe the other voice. Your call.
How loud is the whisper?
Quiet enough that only you hear it. Adjustable in-app, bone-conduction-safe at default. Engineered to land in the gap between heartbeats — felt more than heard.
What if it gets it wrong?
You ignore it. A bad cue costs you nothing — it’s a private suggestion in your ear, never your mouth. Worst case is silence. Best case is the perfect line at the perfect moment. The downside is zero.
Does it work without internet?
Mostly. Whisper STT and the trigger detector run entirely on-device. Only the cue request itself needs network — a few hundred bytes of extracted context, answered in under a second.
What happens after the trial?
If you don’t subscribe, the app stays installed but stops generating new whispers. No auto-charge surprise — we only charge a card if you choose to continue. The smartest twenty dollars you’ll ever spend, but you decide.
When can I actually use it?
Closed TestFlight is live now (Phase 2 · 500 users). Public TestFlight opens at launch (Phase 3 · 5,000 users). Start the trial and we’ll send you the build the moment it goes live.