Why "GenSort" Was Built Without the Cloud
You shouldn't have to surrender your personal library to a server just to find a duplicate image. We built GenSort as a local-first utility because the right to privacy should not be the toll for the right to organization.
You shouldn't have to surrender your personal library to a server just to find a duplicate image. We built GenSort as a local-first utility because the right to privacy should not be the toll for the right to organization.
The modern software ecosystem has trained us to believe that "powerful" implies "connected." If an application performs heavy computation, we assume it requires an array of cloud servers and continuous telemetry. But this is a choice, not a technical inevitability.
When we engineered the visual processing engine for GenSort, the mandate was absolute isolation. The AI models run entirely on your local silicon.
Privacy is not a feature you bolt on; it is an architectural constraint you build around.
By refusing the cloud, we were forced to write more efficient code, optimize memory usage, and respect the user's hardware. The result is a tool that is not only private, but fundamentally faster, as it is completely free from network latency.