Immersive video with location-anchored content

The challenge

The client had a strong idea with a technical unknown at its core: nobody knew whether interactive content could stay anchored to places while the video plays, without sacrificing footage quality. Before building a product, we had to prove the heart of it would hold.

How we worked

• A first phase dedicated only to feasibility: a prototype answering the technical question, with a success criterion agreed before writing any code.
• A benchmark of existing immersive-video and geolocated AR products, to see where the product could stand out and what wasn’t worth reinventing.
• A modular roadmap with priorities set together with the client: first the end-user experience, then the tools for content producers, finally the sales channel for businesses.
• A demo at every release and a shared decision log, so every technical choice had a traceable reason even months later.

Architecture and choices

• Four separate platforms (consumer app, production tools, business portal, backend) instead of a monolith, so different teams and release cycles never block each other.
• A fully serverless backend: during launch the fixed infrastructure cost is close to zero and grows only with real users.
• The same projection model implemented in two engines, one for the apps and one for the web, keeping the experience consistent on every device.
• Content delivered via CDN with adaptive streaming, holding high quality even on mobile connections.

External services, and why these

Outcome

Feasibility proven, apps live on the stores and platforms in production. The project was handed over with technical documentation written for whoever takes it over next: architecture, decision history and open questions.

Stack

• React Native
• Unity
• Three.js
• Next.js
• AWS Lambda
• DynamoDB
• CloudFront
• Cognito
• MediaConvert
• Stripe
• Mapbox