Skip to content

Runbook

QuickTalk Local Adapter

The repository includes apps/cli/quicktalk_bench.py for measuring QuickTalk local adapter load time, first frame, render throughput, and mux time.

Terminal
source .venv/bin/activate
python apps/cli/quicktalk_bench.py \
  --asset-root /path/to/quicktalk/assets \
  --template-video /path/to/template.mp4 \
  --audio /path/to/input.wav \
  --output outputs/benchmarks/quicktalk-output.mp4 \
  --device cuda:0

The JSON output includes:

  • init_seconds
  • audio_feature_seconds
  • first_frame_seconds
  • render_seconds
  • render_fps
  • mux_seconds

OpenTalking End-to-End Path

End-to-end testing should pin the model, TTS provider, and input audio before collecting browser, API, and Worker logs.

Terminal
curl -fsS http://127.0.0.1:8000/health
curl -fsS http://127.0.0.1:8000/models | jq

Record:

  • OpenTalking commit, config file, and non-secret .env settings.
  • Hardware and driver versions.
  • Selected avatar_id, model, and backend.
  • Input audio duration, sample rate, and prompt text.
  • First token, TTS first packet, avatar first frame, browser first frame, and A/V sync.

External Model Services

OmniRT, FlashHead direct WebSocket, or other model services should be benchmarked with their own tools. OpenTalking documentation should reference those results and record only the OpenTalking-side call, queue, and playback behavior.