System Requirements#

Hardware Requirements#

CPU: x86_64. A recent Intel Core Ultra processor is recommended. Older Intel Core / Xeon processors will run the service but may be slower on OpenVINO inference paths.
Memory: 16 GB RAM minimum. 32 GB recommended when using larger TTS models (for example Qwen variants) or when keeping multiple sessions warm.
Disk: 20 GB free SSD space recommended for model assets, the Hugging Face cache, and per-session storage. NVMe is preferred for faster first-run model download and conversion.
GPU (optional): Intel integrated GPU or a supported discrete GPU exposed via /dev/dri for the OpenVINO GPU device path.
NPU: not currently supported by any TTS model in this service. See Configuration > NPU for details.

Device	Minimum	Recommended
CPU	x86_64	Recent Intel Core Ultra processor
Memory	16 GB RAM	32 GB RAM
Disk	20 GB free SSD space	NVMe storage
GPU (optional)	Not applicable	Intel integrated GPU or a supported discrete GPU exposed via `/dev/dri`
NPU	Not supported	Not supported (see Configuration > NPU)

Ubuntu 22.04 LTS (validated) or a compatible Linux distribution with a recent kernel.
For container deployment: Docker Engine and Docker Compose v2.
For GPU acceleration on Linux: Intel/OpenVINO host GPU runtime (e.g. intel-opencl-icd, level-zero) installed on the host. This is a separate prerequisite from the Python dependencies.

The standalone path additionally requires:

sudo apt-get update
sudo apt-get install -y libsndfile1

Outbound internet access on first run to download model assets from Hugging Face, unless models are pre-staged under models/ and the cache.
Inbound access to TCP port 8011 (default) for API clients.