Real-Time Injection QA Webapp

I’m putting together a fully-featured, production-ready web application that will inspect injection-moulded plastic parts in real time by analysing a live camera feed through a Vision API. The system must run on a Windows machine yet be accessible from any browser via a clean, responsive dashboard.

Core workflow
The application will capture each part as it comes off the moulding machine, push the image frames to the Vision API, and instantly decide pass / fail according to defect thresholds I can configure on-screen. Every inspection, result, and high-resolution frame should be logged so I can audit trends or trace specific batches later.

Operator dashboard
I want a web-based dashboard that:
• Streams the live camera view with colour-coded status (pass, fail, uncertain).
• Displays current line statistics (cycle count, yield, defect categories).
• Lets me dive into historical data, filter by job, shift, or cavity, and export CSV or images.

System requirements
• Windows-compatible service or container that auto-starts and recovers after reboot.
• Low-latency communication with the Vision API (choice of API is open if you can justify accuracy and speed).
• Modular design so I can swap cameras or retrain models without rewriting the UI.
• Clear documentation for installation, configuration, and retraining.

Deliverables
1. Source code with build instructions.
2. Packaged installer or Docker image for Windows.
3. Database schema and sample data.
4. User and admin documentation (PDF / markdown).
5. A brief test report proving the system meets throughput, latency, and accuracy targets on my sample set.

Timing & collaboration
There’s no hard deadline, so we can schedule sensible milestones for proof-of-concept, alpha field test, and final production rollout. Quality, reliability, and maintainability matter more to me than rushing.

If you have solid experience with machine vision, real-time streaming, and Windows-hosted web apps, let’s talk technical details and plan the next steps.