Label Reader turns a photo of a pill bottle or prescription form into a short, plain-language explanation anyone can understand — what the medication is for, how to take it, common side effects, and any warnings printed on the label, delivered in the language the user picks (English, Spanish, French, or Portuguese) and available as a read-aloud.
Why we built it: Prescription labels are dense, clinical, and printed in tiny text — hardest to read for exactly the people who most need to get it right: elderly patients, non-native speakers, and caregivers managing someone else's medication. We wanted a tool that anyone could point a phone camera at and instantly understand, without needing to decode medical shorthand or squint at fine print.
The problem it solves:Medication misuse from misread labels is a real, everyday risk. Label Reader adds a built-in safety layer — every scan is flagged as routine , caution, or urgent — so anything unusual is easy to spot at a glance and prompts a check-in with a pharmacist or doctor exactly when it matters, without ever trying to replace professional medical advice.
How it works: Users can upload multiple photos in one go (e.g. front and back of a bottle, or a multi-page form), which are sent to the Gemini API with a structured response schema so the model returns clean, consistent data — medication name, purpose, dosage instructions, side effects, warnings, a safety level, and the raw extracted text. The app is a lightweight Flask service, containerized and deployed on Google Cloud Run, which scales to zero when idle and back up on demand — keeping hosting free for typical usage while still handling traffic spikes during a demo.
