Fix Real-Time Face Swapping App

Job ID: 39749286

Budget: $250 – $750 USD

I need an experienced web developer to fix a real-time face swapping web application that's partially working but has critical video streaming issues.

Current Setup (Already Done)

- ✅ Deep Live Cam and DeepLiveWeb repositories cloned on RunPod GPU server
- ✅ All dependencies installed (Python, ONNX, InsightFace, aiortc, etc.)
- ✅ Face swapping models downloaded (inswapper_128_fp16.onnx, GFPGANv1.4.pth)
- ✅ WebSocket server running on port 5001 with GPU acceleration
- ✅ Frontend served on port 5000 via RunPod proxy
- ✅ WebSocket connections working: wss://fg8gi2ynvdhgge-5001.proxy.runpod.net/
- ✅ WebRTC peer connection establishing successfully

The Problem

Everything connects but no face-swapped video displays:
- Local webcam video shows correctly
- WebSocket connects: "Connected to signaling server"
- WebRTC establishes: "WebRTC connection established"
- BUT: "Remote Video (Swapped)" section stays black
- Face swapping models are loaded and SwapTrack is initialized
- Server processes video frames but processed video doesn't reach browser

Technical Details

- Environment: RunPod GPU server (CUDA 12.4)
- Backend: Python WebSocket + WebRTC (aiortc library)
- Frontend: HTML5 + JavaScript WebRTC implementation
- Repositories:
- https://github.com/hacksider/Deep-Live-Cam (base face swapping)
- https://github.com/lukasdobbbles/DeepLiveWeb (web interface)
- Access: SSH access to RunPod server provided

What's Been Tried

- WebRTC iceTransportPolicy configuration fixes
- CUDA library compatibility fixes
- WebSocket URL corrections for proxy environment
- Debug logging throughout video pipeline
- Server restart troubleshooting

What I Need

An experienced developer who can:
1. Diagnose the WebRTC video streaming issue - why processed video isn't reaching
browser
2. Fix the video pipeline so face-swapped video displays in real-time
3. Work with existing setup (don't rebuild from scratch)
4. Handle RunPod proxy environment challenges if needed

Required Skills

- WebRTC expertise (peer connections, media streams, ICE handling)
- Python backend (asyncio, websockets, aiortc)
- JavaScript frontend (MediaStream API, video elements)
- Video processing pipelines understanding
- Remote server debugging via SSH
- GPU/CUDA environments familiarity