Weather of My Life

All apps

Weather of My Life

MATCHING 1,284 PHOTOS TO THE SKYTurning your photos over...

Your photos remember the weather. Now you can too.

Weather of My Life matches every photo in your library to the real historical weather on that day and place, then shows you the kind of day you are happiest in. Then it tells you when the next one is coming.

Sample library . illustrative weather

Only dates and places

"Weather of My Life" would like to read the date and location of your photos.

The photos themselves stay on your phone. A small on-device model reads smiles for the happiness score, and that score never leaves the device.

Tomi's year . Oct 2025 to Sep 2026

Every photo, with its sky

Sample photos . weather values illustrative

Your weather, decoded

What your best days had in common

365 days

The year as a heat map

Tap any day.

Your year, by weather

The film

0:00 / 0:48

Next 7 days . Manchester

Your kind of day is coming

Sample forecast

Weather alerts, for you only

Not "rain tomorrow". Alerts like "Saturday matches 9 of your 10 best days". One a week at most.

Weather streak

You have used 4 of your last 5 "kind of days" for something you love. Missed one: 29 Aug, spent indoors (you took 0 photos).

The grand mission

Know your weather. Plan your life around it.

Everyone has a kind of day that suits them. Almost nobody knows theirs, so the best days slip past spent indoors. We want every person to know their weather and catch more of it.

Year 1

Photo history plus free historical weather: your best-weather profile, heat map, film, and a weekly "your kind of day" alert.

Year 2

Travel by feel: pick trips and dates by how well a city's climate matches your best days, month by month. Couples see the overlap of their two weathers.

Year 3

A personal climate layer for any calendar: the best day this month for the hike, the wedding photos, the hard conversation. Built on ten years of your own sky.

Blueprint

Why this is new

Photo apps make "memories" slideshows and year recaps by place and face. Weather apps show forecasts and history for a city. Nobody joins your own photo history with what the sky was doing on each of those days to find your personal patterns, and then uses the forecast to point you at your next good day.

The data it runs on

Photos: the capture date and GPS location from each photo's metadata (EXIF), read on device through the phone's photo library permission. Photos never leave the phone.
Weather: Open-Meteo's free Historical Weather API (archive-api.open-meteo.com), which returns daily max and min temperature, precipitation, cloud cover and wind for any coordinates back to 1940, with no API key needed for non-commercial use. Requests are made per city per day range, not per photo, and locations are rounded to about 10 km first.
Limits and gaps: many photos have no location (screenshots, downloads, shared images); the city is then guessed from nearby photos. Reanalysis data is gridded, so a valley or coast can differ from the cell. Smiles are only a proxy for happiness and miss people who rarely photograph themselves. Commercial use of Open-Meteo needs its paid plan.
Fallback: without smiles, "best days" use how many photos you took that day. Without locations, the home city is assumed.

1. The mission

Help every person learn the kind of day they are happiest in, from their own history, and catch more of those days.

2. Why people can't resist it

A fact about yourself you never knew. "You are an 18 to 22 C overcast person" is surprising and sticky.
Nostalgia with a twist. Your own photos, replayed as a weather film, feel new.
A reason to act. "Saturday looks like your kind of day" is a nudge with a deadline.

3. How it works for real

Index. Read capture date and coordinates for every photo on device; group into photo-days per place.
Weather join. Batched calls to the Open-Meteo archive (one per city per date range), cached locally.
Mood signal. An on-device face and smile model (Apple Vision or ML Kit) scores selfies and group photos. Optional, and private.
Patterns. Plain statistics: average mood by temperature band and sky type, minimum sample sizes, confidence shown.
Words and film. One summary call to a fast cheap model (Gemini Flash or DeepSeek via OpenRouter) writes the insights and subtitles; the film is rendered on device from your photos and weather.
Forecast. Open-Meteo's free forecast API scores the next 7 days against your profile.

4. Cost per use

Almost free: the weather data is free and one summary call costs under $0.01 per year of photos.
Free: one year, insights, heat map, weekly alert. Paid ($3 a month or $20 a year): all years, the film in HD, trip planner by climate match, couple overlap.

5. Growth loop

The share card ("I'm a 19 C and grey person") invites "what's mine?". The film is made to be posted. Couples and friends compare weathers and plan trips together, which needs both to install.

6. Risks and guardrails

Privacy: photos and smile scores stay on device. Only rounded coordinates and dates reach the weather API, never linked to an identity.
Faces of others: smile scoring only counts the user's own face once they confirm it; other people are ignored.
Honesty: every insight shows its sample size, and patterns with fewer than 10 photo-days are labelled "early guess".
Wellbeing: no "you were sad this month" framing; insights only point toward good days.

7. MVP build order

On-device photo-day index (date and place) and the Open-Meteo archive join.
Best-weather insight with photo count as the mood proxy, plus the heat map.
Forecast match and the weekly "your kind of day" alert.
Smile scoring, then the film and share card.