I built a local photo search app in a year — 3,500 downloads, zero cloud, zero revenue. Here is what I learned.
My story about building Pitara local Photo search for Windows.
About a year ago I got fed up searching through folders trying to find a specific photo I knew existed somewhere across my external drives. Cloud photo search tools like Google Photos require uploading everything, which was never an option for me. So I built my own.
What it does
Pitara reads the metadata already embedded in every photo — GPS location, date, camera model, altitude — builds a local index using Apache Lucene, and lets you search in plain English. "Summer 2019 in paris", "father's day 2019", "above 8000 feet with canon". Results in under half a second across terabytes of photos.
It runs entirely on Windows, no cloud, no account, no GPU required.
The technical lessons
Lucene.NET is genuinely powerful but the documentation for the .NET wrapper is sparse. I ended up reading Java Lucene docs and translating. Once I understood the schema design the search quality jumped significantly.
EXIF metadata is messier than I expected. Every camera manufacturer handles it slightly differently. iPhone HEIC files have their own quirks. Edge cases took longer to handle than the core search did.
The hardest feature was not the search. It was external drive support. When a drive reconnects with a different letter, the index still has the old paths. Remapping those automatically without requiring the user to re-index took a lot of iteration to get right.
On-device face recognition using ONNX Runtime is very doable without a GPU. It is slower than a cloud API but users with large personal photo libraries actively want their data to stay local. The trade-off is a feature, not a bug.
The distribution lessons
3,500 downloads came almost entirely from organic search and a few Reddit posts. No paid ads, no launch on Product Hunt yet (that is coming this week).
Windows-only turned out to be less of a limitation than I expected. There are very few good offline photo search tools for Windows and that gap drives consistent organic traffic.
Free with an early adopter lock-in model. Everyone who downloads now stays free permanently, even if paid features are introduced later. This has been a good trust signal — no one worries about paying later.
