Rx-to-order · AI

From a prescription photo to a filled cart

An AI that reads a prescription, matches every medicine to your catalogue with stock and cheaper analogues, and builds the order. Built to plug into any pharmacy.

API-firstPrivacy-firstYour own model
Try the live demo Book a pilot
Live prototype · shown in Azerbaijani

A real working prototype. Photograph or upload a prescription and watch it recognise each medicine, match the catalogue, and fill the cart. The market pilot runs in Azerbaijani; we localise the recognition to each pharmacy and language.

How it works

Three taps, no typing

01
Capture
The patient photographs the prescription, or uploads it from the gallery. No forms, no account, no typing of medicine names.
02
Recognise
The model reads every line, handwritten or printed, and matches it to your catalogue with confidence, stock, and cheaper analogues.
03
Order
Confirmed medicines drop into the cart with clear savings and one checkout. Out-of-stock items are swapped for an in-stock alternative.
Why it is different

Built to be embedded, and to forget

API-first
Every capability is an endpoint

Prescription-to-order is a single REST call or SDK. Drop it into your existing app, web checkout, or call centre, no rebuild of your stack.

Privacy-first
Prescriptions are never stored

Images are processed in memory and the response is encrypted in transit. Nothing about a patient's prescription is retained after the order is built.

Your own model
Tuned to your catalogue, not a generic service

The recognition is trained on pharmacy data and tuned to your medicines, your spellings, your stock, so accuracy climbs with every prescription it sees.

Fast
Two to five seconds, photo to cart

Recognition, catalogue match, stock check and analogue suggestion all happen in one pass, fast enough to feel instant at the counter or in the app.

Integration

Three ways to ship it

REST API
One authenticated endpoint. Send an image, receive matched medicines, stock and analogues as JSON.
SDK
A drop-in package for mobile and web, with the capture, scan and cart screens ready to brand.
On-prem
Run it inside your own infrastructure when data must never leave the building.
POST /v1/prescriptions:read
Content-Type: multipart/form-data

{
  "medications": [
    { "name": "...", "dose": "...", "confidence": "high",
      "match": { "sku": "...", "in_stock": true, "analogue": { ... } } }
  ]
}
Questions

What pharmacies ask

How does the prescription scanner work?

The patient photographs the prescription, and the model reads every line, handwritten or printed, then matches it to your catalogue with stock and cheaper analogues in two to five seconds. No typing of medicine names.

Is patient prescription data stored?

No. Images are processed in memory and the response is encrypted in transit. Nothing about a patient's prescription is retained after the order is built.

Can it integrate with our existing pharmacy app or website?

Yes. It is API-first: a single authenticated REST endpoint or an SDK drops into your app, web checkout, or call centre without rebuilding your stack. On-premise deployment is available when data must stay inside your infrastructure.

Does it work with handwritten prescriptions?

Yes. Recognition handles handwritten and printed prescriptions, returns a confidence signal for each line, and flags unclear names for a quick verify step.

Which languages and catalogues are supported?

The recognition is trained on pharmacy data and tuned to your own catalogue, spellings, and stock, not a generic third-party service. The current market pilot runs in Azerbaijani; we localise per pharmacy and region.

Want it in your pharmacy?

We run a short pilot, tuned to your catalogue and your language. No third-party dependency, the model becomes yours.

Start a pilot
← All work