Antalia-2 Mini

Use it to synthesize Turkish speech from text locally through Python, a command-line interface, or the browser demo.

Availability

Install with pip install antalia-mini; browser demo available on Hugging Face Spaces

Pricing model

Not established

Deployment

Runs locally; browser demo uses ONNX Runtime Web on-device with no server

Privacy

Browser demo keeps text on the device; text never leaves the page

Licence

Apache-2.0 for code and model weights

• The model synthesizes Turkish speech from raw Turkish text and has one synthetic male voice. ^[1]
• It produces audio at 48 kHz and contains 7.62 million parameters. ^[1]

  (acoustic model 3.69M, vocoder 3.93M), produces 48 kHz audio, and runs well faster than real time on a

• The package can be installed with pip install antalia-mini, and a browser demo is available on Hugging Face Spaces. ^[1]
• The browser demo runs entirely on the device with ONNX Runtime Web and no server. ^[1]

  runs this release entirely on your device with ONNX Runtime Web (no server; the text never leaves the page).

• Text entered in the browser demo never leaves the page. ^[1]

  runs this release entirely on your device with ONNX Runtime Web (no server; the text never leaves the page).

• The code and model weights are licensed under Apache-2.0. ^[1]

  The code and the weights are licensed under Apache-2.0.

• Turn a prompt into an editable slide, poster, or dashboard Weekly brief 2026-W41
• Turn documents, charts, audio, and video into structured information Daily brief 2026-10-09