SachAI | AI Video Authenticity Platform with .NET & React

SachAI — AI Video Authenticity Platform

Problem Video authenticity analysis needs more than an upload form: long-running inference must handle failures, preserve results and give users a clear view of progress and evidence.

My role I independently designed, developed and launched the product, covering the .NET backend, React frontend and Python AI inference service.

Architecture ASP.NET Core (.NET 10) with Clean Architecture layers (Domain, Application, Infrastructure, API), EF Core/PostgreSQL, a React 19/TypeScript client, and Python FastAPI/PyTorch inference.

What I implemented • Hangfire analysis pipeline with FFmpeg frame/metadata extraction, AI scoring, evidence generation and downloadable PDF reports • Progress tracking, pause/resume, retries, timeouts, a provider circuit breaker, and recovery of stuck or duplicate jobs • Identity, JWT/refresh tokens, Google/Facebook sign-in, owner-based authorization, rate limiting, account safeguards, encryption and audit logs • Versioned EF Core migrations, performance/hash-lookup indexes and perceptual frame hashing for re-upload matching • Cloudflare R2 storage, retention cleanup, subscription/trial limits, scan reservations and reconciliation, Easypaisa integration and an admin dashboard • Serilog with correlation IDs, dependency health checks, Docker, deployment/smoke-test scripts and automated xUnit, pytest, Vitest and Playwright tests • Mock/real AI modes with Hugging Face/PyTorch models and a calibrated ensemble of video- and frame-level detectors

Engineering challenge Coordinating API, storage, worker and inference failures without leaving jobs or reservations in inconsistent states. I built recovery, timeout and reconciliation workflows around these boundaries.

Outcome A live product that connects uploads to asynchronous analysis, stored evidence and reports, with operational tooling and tests. AI scores support assessment; this case study makes no guaranteed accuracy claim.

Live product: https://sachaitech.com/ Backend: https://github.com/KiramatUllah171/ai-video-detection-backend Frontend: https://github.com/KiramatUllah171/ai-video-detection-frontend Python AI service: https://github.com/KiramatUllah171/ai-video-detection-ai-service

Need a .NET API, React integration or reliable background workflow? Message me with your requirements.