How much does it cost to build an app like Amwell?

There is no one figure, because "an app like Amwell" covers three very different builds: a first version with only the essential features, a complete and credible competitor, and a rebuild of everything the mature product does at the traffic it carries. Those are several-fold apart. What this page gives you is the shape of each — the sides, the feature tiers, the stack and the compliance — and the calculator turns whichever one you actually mean into a cost range, hours and a timeline.

Because it is not one app. It is 4 separate applications — patient experience, clinician workstation, clinical cart & device mode, customer configuration console — sharing one backend, each with its own design, release cycle and test matrix. Most quotes that come in low have counted one of them. Configuration is the product: Multi-tenant configurability of branding, workflow, routing and consent is where the engineering goes. Each new toggle multiplies the test matrix, and there is no shortcut — customers buy precisely because they can differ.

Yes, and you should. A first version ships the essential feature tier rather than all three, on fewer surfaces, at a launch-sized audience — materially less than the full product, and the calculator will show you by how much. What it does not cut is compliance: healthcare obligations apply from the first user, so those stay in whatever else is deferred.

It depends on the same three things the cost does — which tier of the product you mean, how many surfaces you ship, and how settled the scope is. The calculator returns a timeline alongside the cost, derived the same way: hours divided across a realistic team shape, never a target date worked backwards from.

Cloud infrastructure, third-party services and annual maintenance, and the calculator reports all three separately from the build. They are kept out of the build figure deliberately: they are operating expenditure rather than capital, and adding the two together produces a number that means nothing. Maintenance in particular is not optional — an app that receives none stops working within about a year as OS releases and SDK deprecations accumulate.

This page carries no figures at all, and nobody outside the company has real ones. What it describes is the product SHAPE, observed from what any user can see — the sides, the feature tiers, the stack, the compliance. Nothing here describes what Amwell actually spent, earns, employs or runs on. The calculator prices building something of that shape today, for the United States market, at our own blended delivery rate.

Only if you handle protected health information as a covered entity or on behalf of one. A fitness tracker you sell directly to consumers generally is not in scope. A patient portal for a clinic, an app a provider gives to patients, or anything where you sign a business associate agreement absolutely is. The distinction is about your relationship with the provider, not the type of data — and it should be settled before you design your data model, because retrofitting HIPAA controls costs several times what building with them costs.

A single HL7 v2 interface through an integration engine is $30,000–$70,000. A FHIR R4 API layer is $45,000–$90,000. An Epic or Oracle Health app-program integration runs $60,000–$160,000 and is gated by their onboarding and review timeline rather than your engineering speed, so it belongs on the critical path from day one. Read-only access is dramatically cheaper than write-back, and starting read-only is usually the right call.