Camar is an open-source speech AI project providing text-to-speech and speech-to-text capabilities for the Turkmen language. It features custom architectures like VITS for synthesis and Whisper-based models for recognition, optimized for local deployment with limited hardware resources.