Activity and mood
See the latest observed activity, mood and confidence, with the model and observation time.
Applet library · template 1.1.0
Pet monitor
A little more peace of mind.
Turn a spare Android device into a quiet observer. See what your pet has been doing, with useful daily patterns and a private place to check in.
Activity · Sleep · Meals · DrinksCapabilities, not live readings
What it can do
Sleep and naps
Follow observed sleep, nap counts and the current sleep session.
Meals and drinks
Count observed eating and drinking episodes, see the latest sightings, and compare observed gaps.
Daily patterns
Understand monitoring coverage, sampled photos and activity distribution. Unchanged frames are labeled as reused.
A private photo diary
With explicit personal-backend and photo consent, browse date-grouped photos and their observations. Up to 100 photos are retained for seven days.
Invite someone you trust
Explicitly invite a viewer to this applet on this device. They can view its images and metrics; you can revoke access.
The current vision classifier recognizes dogs. Observations describe what was seen while monitoring, not everything your pet did all day.
Make room for a new routine.
Choose a phone camera or a local network camera assigned to this applet. Keep the device powered and permit the camera access you want.
Your device. Your choice.
Device only is the default. Hosted AI requires explicit consent for this applet on this device; a local classifier may be unavailable. Saving a template does not start monitoring or enable uploads.
A personal backend is optional for each applet on each device. Invitations grant read-only access to that specific instance and can be revoked.
On the device, open Apps → Pet monitor to configure and start monitoring. My apps displays instances whose personal backend you explicitly enabled. Customizing a template saves a definition; it does not install or start an applet.
Metric definitions and template reference
This page is generated from the same versioned definition seeded into the authenticated template API. Public capability catalog · Full authenticated definition
- Current observation (observation)
- Latest capture: dog, confidence, activity, mood, model and timestamp. Unknown confidence remains null.
- Observed sleep (ms)
- Sum classified sleep intervals during coverage today; sleep episode gap 120000 ms. Never imply continuous observation through a coverage gap.
- Naps (episodes)
- Observed sleep episodes today; first-to-last observation plus one sampling interval.
- Current sleep start (timestamp)
- Start of ongoing observed sleep episode; null when unavailable.
- Current sleep duration (ms)
- Current episode span, may include brief gaps; separate from classified sleep total.
- Observed meals (episodes)
- Observed eating episodes today, merged within 60000 ms.
- Observed drinks (episodes)
- Observed drinking episodes today, merged within 60000 ms.
- Last observed meal (timestamp)
- Latest observed eating timestamp; missing is not zero.
- Last observed drink (timestamp)
- Latest observed drinking timestamp; missing is not zero.
- Monitoring coverage (ms)
- Sum camera monitoring spans, no credit across gaps longer than the runtime coverage threshold.
- Sampled photos (photos)
- Number of sampled captures today, including reused classifications.
- Activity distribution (photo_counts)
- Counts of photos by activity including no_dog and unsure. Counts are not inferred durations.
- Meal gap median (ms_and_gaps)
- Seven-day median of qualifying observed episode gaps within coverage, accompanied by gap count. Null if insufficient data.
- Drink gap median (ms_and_gaps)
- Seven-day median of qualifying observed drinking gaps within coverage, accompanied by gap count.
- Reused frames (photos)
- Unchanged captures reusing a prior classification, labeled reused; never count as model calls.
- Model calls (calls)
- Actual processing calls, distinct from sampled photos and reused frames. Do not infer a monetary price.