Thaana AI launches voice cloning and new AI voices

Thaana AI has introduced voice cloning and new text-to-speech models for natural, expressive Dhivehi and English audio.

The update supports Dhinglish within a single sentence, pronunciation tailored to Maldivian names, and adjustable pace and pitch.

According to Thaana AI, the models support the way Dhivehi and English are used together in everyday conversation.

Expressive, human-like voices bring emotional nuance to storytelling, educational material and professional communication.

Before generating audio, the system uses text normalization to expand abbreviations, interpret numbers and resolve ambiguities in context. This helps names, dates, times and acronyms sound clear within a sentence.

For example, a Dhinglish appointment reminder can combine Thaana script with English names, abbreviations and numbers:

‪Dr. Aishath އާ ‪appointment އޮތީ ‪18 Jan ގެ ‪7:30 AM ގައި.

The system is designed to read “Dr.” as “Doctor”, interpret the date and time in context, and pronounce the name naturally as the script moves between Dhivehi and English.

Voice cloning lets users generate narration using an AI version of their own voice, maintaining a familiar sound across videos, storytelling and regular audio content.

Users can assign different voices to sections of one script, creating distinct characters or speakers for dialogue, storytelling and promotional videos. Adjustable pace and pitch provide further control over delivery.

Creators and businesses can produce narration for YouTube videos, podcasts, audiobooks, product demonstrations, advertisements and presentations. Teachers can turn lessons and revision materials into audio, while organizations can narrate staff training resources. Government offices can create spoken versions of public information and service explainers.

Thaana AI is also launching a text-to-speech API that lets organizations add spoken narration to their websites. According to the company, the integration can support accessibility for people with visual impairments or reading difficulties and offer visitors an alternative way to access written content.

Voice cloning also raises questions about consent, impersonation and the authenticity of recordings used as digital evidence.

Addressing this, Thaana AI has released an audio verification tool at thaana.ai/verify-audio to help users check whether audio was generated on its platform, as one component of wider verification and forensic review.

The company called on the Maldivian government and relevant agencies to establish a working group with law enforcement, the legal community, independent forensic experts and AI developers, including Thaana AI US-based Synetecs. It advocates a formal role for innovators in strengthening digital-evidence authentication, developing independently tested safeguards and building local technical capacity, while protecting public trust and supporting innovation in Dhivehi AI.

Synetecs, the deep tech company behind Thaana AI, was founded by Maldivian entrepreneurs Dharis, Ismail and Jinas with the aim of making local AI affordable and expand into global markets.