Natiq Studio has released Natiq-dz-ARG, an open-weight text-to-speech (TTS) model specifically designed for Algerian Darija, addressing the limitations of mainstream TTS systems that struggle with code-switching between Arabic, French, and Amazigh influences. The model is built by fine-tuning the 3.4B-parameter Higgs TTS foundation model (bosonai/higgs-tts-3-4b) using Low-Rank Adaptation (LoRA), which enables efficient adaptation by training only a small set of additional parameters. Training utilized over 300 hours of speech data from more than 100 diverse speakers to prevent overfitting and improve generalization across voices and accents. A key focus was handling mixed-language transcripts, where French words are often phonetically transliterated into Arabic script, requiring careful data cleaning to preserve linguistic accuracy. Natiq-dz-ARG supports 24 kHz audio output and includes zero-shot voice cloning, allowing synthesis using short reference audio clips. The model is publicly available on Hugging Face under the CC BY-NC 4.0 license for non-commercial use, with commercial licensing requiring separate agreements. Additionally, Natiq Studio offers a hosted platform with a REST API for generating Algerian Darija voiceovers, featuring multiple AI voices and downloadable WAV files, targeting creators and developers building localized applications.