System Live-Flow Monitor

ENGINE_ONLINE | RTC_ACTIVE

Machine Specs

CPU: 4+ Cores

GPU VRAM: 6GB NVIDIA

CUDA: 12.1+

RAM: 16GB DDR4

AI Model Stats

Whisper: 460 MB

Qwen LLM: 1.1 GB

ECAPA-TDNN: 80 MB

Enrollment

Voice Fingerprints locked as Ground Truth.

Capture

16kHz stream sliced into micro-blocks.

Processing

Dual-model inference (Who + What).

Summary

Final LLM report synced to frontend.