New · 2.4 releaseListening
For cats and dogsFree to try

The quiet language of pets, finally put into words.

CatDog turns barks, meows, growls and trills into plain English — and quietly tracks the moods behind them, day after day, like a notebook your pet is writing for you.

See how it listens
Sounds parsed
3.4M
Mood states
16
Avg. accuracy
94.7%
9:41
REC

Mochi · 14:32

Translation

Detected · cat · 96%

“The bowl is empty and you have not noticed.”

hungryinsultedpatient
What it does

Built around how pets actually communicate.

Four tools, distilled from twelve years of bioacoustic research and forty-seven thousand pet recordings. Not a single bark wasted.

Live translation

A 600ms window from sound to sentence.

Hold to record. The model parses pitch contour, formants, and timing — then writes back what it heard, in your language.

“I would like breakfast.”

Mood states

Sixteen feelings, observed daily.

Beyond happy and sad. The model tracks the small ones too — bored, attentive, suspicious, hopeful — and their shifts across hours.

Content38%
Playful22%
Hungry14%
Curious11%
Suspicious7%
Drowsy5%
Bored2%
Hopeful1%

This morning · 07:42

Content, leaning curious

Behavior journal

Twenty-eight days, in dots.

Every recording adds a colored dot. Patterns surface in a week.

Content Playful Hungry Quiet
Mood trend

Quietly trending up.

This week

+18.4%

Streak

11 days

How it listens

Three motions, then the conversation begins.

01Press the brass button

Hold to listen.

A tap won’t do — the model needs the full breath of a sound to catch its shape. Hold while your pet finishes the thought.

DetailAvg. recording: 1.6s
02Translation arrives

Read the line.

Within a second the line appears, plain-spoken. You’ll get the most likely interpretation, with two alternates if it was close.

DetailLatency: 600–900ms
03Or just let it pass

Save the moment.

Worth keeping? File it to the journal with one tap. Otherwise, it dissolves and the model never holds your audio.

DetailAudio retention: 0s

Nothing leaves the device unless you ask it to. Audio is processed on-phone, journaled where you choose, deleted on your schedule.

Inside the model

Your pet has roughly one hundred things to say.

The model was trained on 47,000 annotated recordings across 31 breeds. It listens for pitch envelope, harmonic ratio, and the silences between — the same signals vets and animal behaviorists use, only faster.

Mood categories
16
Avg. accuracy
94.7%
Species supported
Cat / Dog
Process latency
< 1.0s
Audio retention
0s
Source · meow · 1.4sLive
Translation · cat · 96%

attentionmildly annoyed
fig.03 · live capture
From the journals

Two million households, listening a little harder.

Real notes from real pet parents, shared with permission.
I thought she was being needy. Turns out she was bored at 4pm every day and I was missing it. We rearranged the schedule and she stopped knocking things off counters.
MPMargaux PelletierCat parent · 7 months in

The journal caught Captain’s anxious days clustering around trash pickup. We moved the bin. I would not have spotted that pattern in a hundred years.

RERafael EstévezDog parent · 14 months in
CompanionCaptain / Labrador

My vet asked what changed. I showed her the mood graph. She wants to see it for every patient now.

TATomoko AraiShiba parent

It does not gaslight you about your pet. It just tells you what they said and lets you decide what to do.

HOHannah OkonkwoTwo cats · 9 months in
Modern CatThe BarkApartment TherapyWired
Get the app

Hold the button. Listen for a year.

Free to download. No account required to try.

iPhone only · for now
Account
Optional, anonymous
Audio
Processed on-device
Device
iPhone, iOS 17+
cat/dog · made with whiskersAudio processed on-device