Backend Modes

OpenTalking keeps orchestration, LLM, STT, TTS, avatar management, and WebRTC in the OpenTalking process. The talking-head runtime is selected through a backend. The deployment docs focus on local and omnirt.

Mode	Process shape	Use it when	Cost
local	OpenTalking imports and runs the model adapter in-process.	Single-machine validation, minimal components, fast debugging.	Model dependencies and VRAM share the OpenTalking process.
omnirt	OpenTalking connects to an external OmniRT service.	Service isolation, model gateways, GPU/NPU-specific runtime.	Requires a separate OmniRT checkout, .venv, port, and model service.

Recommended Layout

Terminal

export DIGITAL_HUMAN_HOME="$HOME/digital-human"
export OPENTALKING_HOME="$DIGITAL_HUMAN_HOME/opentalking"
export OMNIRT_REPO="$DIGITAL_HUMAN_HOME/omnirt"
export OMNIRT_HOME="$OMNIRT_REPO/.omnirt"
export OPENTALKING_MODEL_ROOT="$DIGITAL_HUMAN_HOME/models"
export OMNIRT_MODEL_ROOT="$DIGITAL_HUMAN_HOME/models"

Package Mirrors

Terminal

export UV_DEFAULT_INDEX="https://pypi.tuna.tsinghua.edu.cn/simple"
export PIP_INDEX_URL="https://pypi.tuna.tsinghua.edu.cn/simple"
export UV_HTTP_TIMEOUT=300
export UV_LINK_MODE=copy

Keep UV_LINK_MODE=copy when the uv cache and .venv live on different filesystems; it avoids cross-device hardlink fallbacks leaving dependency installs in a bad state.

Verification

Terminal

curl -fsS http://127.0.0.1:8000/health
curl -s http://127.0.0.1:8000/models | python3 -m json.tool
curl -fsS http://127.0.0.1:9000/v1/audio2video/models | python3 -m json.tool

Next Pages

• Local Adapter
• OmniRT
• Talking-Head Models
• Support Matrix

Frontend Entry

After the model or backend service is running, use the OpenTalking WebUI:

Terminal

cd "$OPENTALKING_HOME"
bash scripts/quickstart/start_frontend.sh --api-port 8000 --web-port 5173 --host 0.0.0.0

For a remote server, forward your local browser port to the server 5173, then open http://127.0.0.1:5173.