Honest answers about matching, accuracy, privacy, costs, what happens to your selfie, and how photographers get paid. If your question is not here, the contact page reaches us directly.
You take one selfie. Our system sends it to AWS Rekognition, which detects your face and turns it into a feature vector (roughly 128 numbers describing the geometry of your face, like spacing between eyes and curve of the jaw). That vector is searched against every face indexed from the event you came to find. Anything within the similarity threshold comes back as a match. The whole pipeline runs in about two to five seconds.
For frames where you are roughly facing the camera and your face is at least thumbnail-sized in the original full-resolution file (which is much bigger than it looks in the preview), we see correct match rates above 98% at the high-confidence threshold. For profile shots, partial faces, or very small faces in wide crowd shots, recall drops to 60-70%. Outfit matching picks up some of that slack. False positives above the 80% confidence threshold are well under 1%.
An optional second pass. After the face match runs, we look for photos with the same clothing color blob below a face that did not score high enough on face matching alone. Then we re-verify each candidate by running face matching on it again, so we never hand you photos of a stranger wearing a similar jacket. Outfit matches show up with a purple badge, face matches with a blue one.
Your selfie generates a face fingerprint, then the selfie image itself is deleted from our storage unless you explicitly opt in to save it for future events. If you opted in, the selfie sits in a private S3 bucket that is not publicly accessible, and you can delete it any time from the profile screen.
No. Face vectors are not sold to third parties, not shared with advertisers, and not handed off to data brokers. The infrastructure that holds them (AWS Rekognition collections) does not have an export path that ships them to anyone, and we built it that way on purpose.
From the profile screen in the app, go to Account Settings and tap Delete Face Data. That removes your face vectors from every Rekognition collection they are in and deletes any saved selfie images. If you also want to close your account entirely, email hello@initpic.com to delete your account permanently. Once deleted, your face data is gone immediately and cannot be recovered.
Only the ones you actively choose to search. Each event has its own Rekognition collection. When you search a specific event, only that collection is queried. The optional "save my selfie for future events" feature speeds up matching at your next event but still requires you to open that event before searching runs.
Searching is always free. You only pay for the photos you actually want to download or print. Pricing is set per photographer, but most photo events run digital downloads in the $4-6 range with automatic bundle discounts: 10% off at five photos, 20% off at ten.
Festivals, weddings, marathons and races, concerts, corporate events, conferences, brand activations, parties, sports events. Any event where a photographer was hired to shoot the crowd. We have dedicated pages for the most common: festivals, weddings, marathons, concerts, and corporate events.
The original full-resolution file the photographer uploaded, with no watermark. For most modern event photographers that is 24 megapixels or higher, suitable for printing at large sizes.
Photographers receive 85% of every digital photo sale, paid out via Stripe Connect. Print products carry a Printify fulfillment cost plus a configurable platform margin (default 100% markup on Printify cost). Earnings, top events, and a daily sales chart are visible live in the photographer dashboard.
Yes. Photographers upload, review, and publish their galleries themselves. Nothing is auto-listed without their approval. Watermarked previews are generated automatically during upload, but the photographer controls when the event is visible to attendees.
Our chunked upload pipeline handles multi-gigabyte galleries. We routinely process 25,000-frame marathon uploads in one go. Per-image cap is 25MB, with 50 images per request batch and direct-to-S3 chunked uploads for files over 50MB. The full operational guide is in our photographer playbook.
Either the score is in the lower confidence band and the system is hedging, or someone in the frame really does have face geometry close to yours. High-confidence matches are almost always you. Lower ones are worth a quick glance and a thumbs down if they are wrong, which trains the system for next time.
Face matching does not care about outfits. You will still match every photo where your face is visible. Outfit matching is an optional second pass that helps catch frames where your face is turned away. If you changed outfits, two selfies (one per look) maximize recall on the outfit pass.
For most events results come back in two to five seconds. Very large events with hundreds of thousands of indexed faces can take ten to fifteen seconds the first time. After that, your results are cached and reload instantly. If you are stuck longer than thirty seconds, refresh the page and try once more.
One selfie. Every photo of you. No account required to search.