Chariot launches multilingual speech model for 11 Indian languages and English
The AMW Read
The release adds an Indian-language speech model to the generative voice market, while its competitive performance remains based on company-reported benchmarks.
Chariot launches multilingual speech model for 11 Indian languages and English
Indian AI lab Chariot has launched Chariot-1-TTS-Multilingual, a text-to-speech model supporting 11 Indian languages and English. It offers more than 50 voices, switches between two supported languages within a sentence, and clones voices from a few seconds of reference audio. Each request can generate speech from up to 1,000 characters. The company says the model uses context to determine how to pronounce written expressions, including dates, currency amounts, and abbreviations.
The launch adds another company to the market for generative voice tools, with a focus on multilingual communication in India. Chariot says the model will be used in voice agents, customer service, education, and public helplines. It reports character error rates of 7.60% for Hindi and 6.47% across nine other Indic languages, as well as 93.78% accuracy on an English verbalization benchmark. These are company-reported results, and the source notes that benchmark coverage varied across competing models.
For builders assessing the model, the practical question is whether its code-mixing, pronunciation, and voice quality hold up in their own languages and workflows. The 1,000-character request limit is also relevant when designing longer speech output. Chariot was selected under the IndiaAI Mission's foundation-model pillar, alongside companies including Sarvam AI and Gnani AI.