Arabic-first expressive zero-shot speech synthesis — Audar-TTS-V1 (Flash + Turbo + Pro). Voice cloning, expression tags, benchmarks, technical report & inference.
