Visual Evidence App
About the project
Many workflows are highly manual from the employee or users perspective. By handing the user a self-service App, the workflow progress can be easily tracked, and a lot of details captured, only to be used if the case needs review. If you scan the QR code of your laptop next to the repair desk, you probably want to drop it off or pick it up.
- Built using modern full-stack technologies (Node.js, Deno, Svelte)
- Implemented secure authentication and session flows
This can be used to create a workflow as a simple webapp using React, Svelte or plain JS and Web Components
- Possible to implement async updates
- Can Filter PII for better privacy and compliance
What we did
With further integrations in mind, we built a scalable and flexible base.
- Deployed using European CDN + serverless infrastructure
- Set up an automated build, test, deployment based on GitHub
- Embeds advanced ML Vision models to determine workflow state
Android is a future possibility

Modern performance
The architecture allows several advantages
- Native camera performance ensures good battery performance and experience
- Private information can be kept natively on users device
- Native ML Models are available
- No streaming to edge hosted ML algos
- Simple web deployment convenience
- Control over data retention
- High efficiency with light-weight web apps