We grew up in the week after the wedding
SpotMe was started in Ahmedabad by two engineers who had watched photographers lose their evenings to sorting photos by hand. We build one thing, getting every photo to the right person, and we try to build it properly.
SpotMe in numbers
- studios deliver weddings with SpotMe
- 650+
- photos matched to the right guest
- 1.2M+
- face-match accuracy on real weddings
- 99.99%
- from selfie to a guest’s own gallery
- ~1 s
Umang Gajjar
Co-founder and CEO, Ahmedabad
“We are not trying to replace the photographer. We are trying to give them their evenings back.”
I grew up carrying reflectors at weddings. My uncle, Rajesh-kaka, ran a small studio in Maninagar, and from the age of twelve my winter holidays were spent at farmhouse receptions and pol-house pithi ceremonies, holding lights and changing film, and later memory cards.
What I remember most is not the weddings. It is the week after. Kaka would sit at his desktop until two in the morning, scrolling through thousands of photos and copying them into folders named after people: “Bhabhi’s side”, “Dhruv friends”, “Nani”. Then the phone calls started, and never really stopped.
I became an engineer instead of a photographer. I spent ten years building software, first in Bengaluru and then in Toronto, mostly products that moved a lot of data quietly in the background. In 2023 I flew home for my cousin’s wedding. Five functions, 7,400 photos, one Google Drive link in the family group. Within a day I was the person everybody called, because I was “good with computers”. I spent my holiday making face folders for relatives, exactly like Kaka used to.
That week Bansal and I sketched the first version of SpotMe on a napkin at a tea stall near Law Garden. The idea was simple: the photographer uploads once, every guest takes a selfie, and the software does what my uncle did by hand, only properly and in a second.
We started in Ahmedabad because this is where the problem was loudest, and where we could sit with photographers in their studios and watch them work. Most of our early customers were friends of Kaka. They were not shy about telling us what was wrong.
I still go to a wedding with a studio at least once a month. It keeps us honest.
Bansal Thakkar
Co-founder and CTO, Ahmedabad
“An accuracy number only matters if you measured it on the photos your customers actually take.”
Umang and I met in our second year of engineering in Ahmedabad, on a hackathon team that lost badly. We have been building things together, on and off, ever since.
I spent most of my career in computer vision, teaching machines to see things in images that are messy, crowded and badly lit. Before SpotMe I worked on medical imaging and on retail analytics, which taught me two things: real-world photos are nothing like the clean pictures in research papers, and accuracy numbers mean nothing until you test them on the exact photos your customers have.
Indian weddings are one of the hardest places to recognise a face. Brides under heavy veils and jewellery. Grooms in safas that hide half their forehead. Sangeet lighting that turns everyone purple. Uncles who are only ever photographed from the side while they dance. Children who change outfits three times a day. Off-the-shelf face recognition, the kind most apps rent from big cloud companies, was never built for this.
So we built our own. We trained SpotMe’s models in-house on the conditions that actually show up at weddings, and we tested them on real events, function by function, until we could trust them with a family’s photos. Today we measure 99.99% accuracy on real weddings, and we keep measuring, because every new season brings a new trend in lighting or makeup.
The other decision I care about is where the models run. They run on our own servers, in India. No photo and no selfie is ever sent to an outside AI company, and selfies are never stored. When a mother hands her phone to a stranger’s software to find photos of her child, that is the least she should expect.
How we got here
- 2023
The cousin’s wedding
7,400 photos, one Drive link, and a week of making face folders by hand. The idea for SpotMe starts at a tea stall near Law Garden.
- 2024
SpotMe is founded in Ahmedabad
The first version goes to twelve studios in Ahmedabad and Vadodara. Most of them are friends of Umang’s uncle, and very honest critics.
- 2025
Our own face recognition
Models trained in-house replace the first prototype. Accuracy on real weddings crosses 99.99%, and couple galleries with a people view launch.
- 2026
650+ studios in three countries
Studios across Gujarat and India, plus Indian-wedding specialists in Canada and the USA, deliver more than a million matched photos.
What we won’t compromise on
Four rules we come back to whenever we decide what to build next.
The photographer’s brand comes first
Your studio name and logo lead every gallery. We stay quiet in the corner.
Originals are sacred
We store the file you uploaded exactly as it is, and guests can always download it at full resolution.
Faces stay private
Our own models, on our own servers. Selfies are never stored and face data is never used for training.
Price the storage, not the photos
Every feature on every plan, no credits, no guest caps. Delete an event and the space comes back.
Come and see us in Ahmedabad
Our team works from Ahmedabad. If you’re in Gujarat, we’re happy to visit your studio, or have you over for chai and a demo on one of your own weddings. Everywhere else, we set up your first event with you on a video call.
SpotMe
A product of HPK AI Labs
Ahmedabad, Gujarat, India
Serving studios in India, Canada and the USA.
Your first wedding, set up with you, free.
Send us one event. Our team sets it up with you on a call, and you see it working on your own wedding before you pay anything.
- No card needed
- Every feature included
- 10 GB free forever