ReadingScan

I built this because people photograph a radiology report, paste the terms into a search engine, and spend the week before their appointment convinced of the worst thing they found. ReadingScan reads their report instead: upload the PDF, or a photo of the X-ray, MRI, CT, ultrasound or mammogram itself, and get a patient-facing report in a few minutes — key takeaways, an urgency level, each finding explained in lay terms with the original medical term kept beside it, what to do next, questions for the appointment, and a glossary of the words in that specific report. Follow-up questions about your own report work in chat.

The conservative part is the point. Uploads pass MIME, size and resolution prechecks, then one multimodal model call runs four ordered steps: form its own reading of the image, classify the study, extract metadata, then write the report or refuse the input. The exam date is overwritten server-side rather than trusted to the model. A selfie, a handwritten clinic note or a blurry photo is refused with an explanation, not interpreted anyway.

English, Spanish, Brazilian Portuguese. Pay per report, no subscription — first report free in most countries, then report passes from $14. It explains what a report says so you can have a better conversation with your clinician. It is not a medical opinion and not for emergencies.