Language Coverage
Maya Research supports 12 languages across two script families: Indic scripts (Devanagari, Tamil, Telugu, Kannada, Malayalam, Bengali, Gujarati, Gurmukhi) and Arabic script (MSA and two spoken dialects). All languages are available on Growth and Pro plans. Free plan includes 2 languages.
Indic Languages
| Language | Native Name | Script | BCP-47 Code | Voices | Dialect Notes | Status |
|---|---|---|---|---|---|---|
| Hindi | हिन्दी | Devanagari | hi-IN |
4 | Standard Hindi. Khari Boli prosody. SSML emphasis and rate supported. | GA |
| Tamil | தமிழ் | Tamil | ta-IN |
3 | Standard Tamil (Chennai register). Classical phoneme support included. | GA |
| Telugu | తెలుగు | Telugu | te-IN |
3 | Standard Telugu. Velar and dental phoneme distinction maintained. | GA |
| Kannada | ಕನ್ನಡ | Kannada | kn-IN |
2 | Standard Kannada. Retroflex consonants modeled explicitly. | GA |
| Malayalam | മലയാളം | Malayalam | ml-IN |
2 | Standard Malayalam. Chillu consonant handling included. | Beta |
| Bengali | বাংলা | Bengali | bn-IN |
2 | Indian Bengali (West Bengal register). Bangladeshi register in beta. | Beta |
| Marathi | मराठी | Devanagari | mr-IN |
2 | Standard Marathi. Shares Devanagari script with Hindi but distinct prosody model. | Beta |
| Gujarati | ગુજરાતી | Gujarati | gu-IN |
2 | Standard Gujarati. Aspirated stop handling modeled accurately. | Beta |
| Punjabi | ਪੰਜਾਬੀ | Gurmukhi | pa-IN |
1 | Indian Punjabi (Gurmukhi script). Tonal contrast (flat, rising, falling) included. | Beta |
Arabic Variants
Maya Research supports three Arabic variants reflecting the major registers in use across the Arab world. Each variant has a distinct prosody model trained on dialect-specific corpora; they are not adaptations of a single MSA model.
| Variant | Native Name | BCP-47 Code | Voices | Use Case | Status |
|---|---|---|---|---|---|
| Modern Standard Arabic | العربية الفصحى | ar-001 |
3 | News, formal announcements, educational content. Understood across the Arab world. | GA |
| Egyptian Arabic | عربي مصري | ar-EG |
2 | Consumer apps, conversational IVR, entertainment. Widely understood as a spoken dialect. | GA |
| Levantine Arabic | عربي شامي | ar-LV |
2 | Syria, Lebanon, Jordan, Palestine regional apps. More limited geographic reach than Egyptian. | Beta |
Language Code Quick Reference
Indic languages
hi-IN Hindi (Devanagari)
ta-IN Tamil
te-IN Telugu
kn-IN Kannada
ml-IN Malayalam
bn-IN Bengali (India)
mr-IN Marathi (Devanagari)
gu-IN Gujarati
pa-IN Punjabi (Gurmukhi)
Arabic variants
ar-001 Modern Standard Arabic
ar-EG Egyptian Arabic
ar-LV Levantine ArabicSSML Support by Language
Maya Research supports a subset of SSML tags. Support varies by language because some tags (especially prosody rate and emphasis) interact with language-specific phonological rules.
| SSML Tag | hi-IN, ta-IN, te-IN, kn-IN | ml-IN, bn-IN, mr-IN, gu-IN, pa-IN | ar-001, ar-EG, ar-LV |
|---|---|---|---|
<break> |
Full support | Full support | Full support |
<prosody rate> |
Full support | Partial (0.8-1.4x) | Full support |
<prosody pitch> |
Full support | Beta | Beta |
<emphasis> |
Full support | Reduced | Not supported |
<say-as> |
Digits, date, phone | Digits only | Digits, date |
<phoneme> |
IPA (experimental) | Not supported | Not supported |
For a deeper discussion of SSML prosody behavior on Indic models, see the Lab article: SSML Prosody Tags for Indic TTS: What Works, What Does Not.