# Language Support

> Configure voice agents for 50+ languages with optimal STT, TTS, and LLM provider combinations.

Source: https://edesy.in/docs/voice-agent/languages

Edesy voice agents support 50+ languages with automatic provider selection optimized for each language.

## Supported Languages

### Tier 1 (Excellent Support)

| Language | Code | STT | TTS | LLM | Notes |
|----------|------|-----|-----|-----|-------|
| English (US) | en-US | All | All | All | Best support |
| English (UK) | en-GB | All | All | All | British accent |
| English (IN) | en-IN | All | Azure | All | Indian accent |
| Spanish | es | All | All | All | Latam + Spain |
| French | fr | All | All | All | France + Canada |
| German | de | All | All | All | |
| Portuguese | pt-BR | All | All | All | Brazil |
| Italian | it | All | All | All | |
| Japanese | ja | All | All | All | |
| Korean | ko | All | All | All | |
| Chinese (Mandarin) | zh-CN | All | All | All | Simplified |

### Tier 2 (Good Support)

| Language | Code | Best STT | Best TTS | Notes |
|----------|------|----------|----------|-------|
| Hindi | hi-IN | Google Chirp | Azure | Excellent support |
| Bengali | bn-IN | Google Chirp | Azure | |
| Tamil | ta-IN | Google Chirp | Azure | |
| Telugu | te-IN | Google Chirp | Azure | |
| Marathi | mr-IN | Google Chirp | Azure | |
| Gujarati | gu-IN | Google Chirp | Azure | |
| Kannada | kn-IN | Google Chirp | Azure | |
| Malayalam | ml-IN | Google Chirp | Azure | |
| Punjabi | pa-IN | Google Chirp | Azure | |
| Dutch | nl | Deepgram | Azure | |
| Polish | pl | Deepgram | Azure | |
| Russian | ru | Deepgram | Azure | |
| Turkish | tr | Deepgram | Azure | |
| Arabic | ar | Google | Azure | MSA + dialects |
| Hebrew | he | Google | Azure | |

### Tier 3 (Basic Support)

| Language | Code | Best STT | Best TTS | Notes |
|----------|------|----------|----------|-------|
| Assamese | as-IN | ElevenLabs | Azure | Limited vocabulary |
| Odia | or-IN | Google | Azure | |
| Nepali | ne | Google | Azure | |
| Sinhala | si | Google | Azure | |
| Thai | th | Google | Azure | |
| Vietnamese | vi | Google | Azure | |
| Indonesian | id | Deepgram | Azure | |
| Malay | ms | Google | Azure | |
| Filipino | fil | Google | Azure | |
| Ukrainian | uk | Google | Azure | |
| Czech | cs | Deepgram | Azure | |
| Romanian | ro | Deepgram | Azure | |
| Hungarian | hu | Deepgram | Azure | |

## Configuration

### Basic Language Setup

```json
{
  "agent": {
    "name": "Hindi Support Agent",
    "language": "hi-IN",
    "llmProvider": "gemini-2.5",
    "sttProvider": "google",
    "sttModel": "chirp_2",
    "ttsProvider": "azure",
    "ttsVoice": "hi-IN-SwaraNeural"
  }
}
```

### Auto-Detection

Enable language detection for multi-lingual support:

```json
{
  "agent": {
    "name": "Multi-lingual Agent",
    "language": "auto",
    "supportedLanguages": ["en-US", "hi-IN", "es"],
    "sttConfig": {
      "enableLanguageDetection": true
    }
  }
}
```

## Provider Recommendations by Language

### English (en-US, en-GB)

```json
{
  "sttProvider": "deepgram",
  "sttModel": "nova-3",
  "ttsProvider": "cartesia",
  "llmProvider": "gemini-2.5"
}
```

**Why**: Deepgram Nova-3 has the lowest latency and highest accuracy for English. Cartesia provides natural-sounding voices with fast streaming.

### Hindi (hi-IN)

```json
{
  "sttProvider": "google",
  "sttModel": "chirp_2",
  "ttsProvider": "azure",
  "ttsVoice": "hi-IN-SwaraNeural",
  "llmProvider": "gemini-2.5"
}
```

**Why**: Google Chirp 2 is specifically optimized for Indic languages. Azure has the best Hindi voices.

### Available Hindi Voices (Azure)

| Voice | Gender | Style |
|-------|--------|-------|
| hi-IN-SwaraNeural | Female | Warm, conversational |
| hi-IN-MadhurNeural | Male | Professional |
| hi-IN-AnanyaNeural | Female | Clear, formal |
| hi-IN-ArjunNeural | Male | Young, friendly |

### Tamil (ta-IN)

```json
{
  "sttProvider": "google",
  "sttModel": "chirp_2",
  "ttsProvider": "azure",
  "ttsVoice": "ta-IN-PallaviNeural",
  "llmProvider": "gemini-2.5"
}
```

### Bengali (bn-IN)

```json
{
  "sttProvider": "google",
  "sttModel": "chirp_2",
  "ttsProvider": "azure",
  "ttsVoice": "bn-IN-TanishaaNeural",
  "llmProvider": "gemini-2.5"
}
```

### Assamese (as-IN)

```json
{
  "sttProvider": "elevenlabs",
  "ttsProvider": "azure",
  "ttsVoice": "as-IN-YashicaNeural",
  "llmProvider": "gemini-2.5"
}
```

**Note**: Assamese has limited STT support. ElevenLabs Scribe provides the best accuracy.

## Language-Specific Configuration

### Script Direction

For RTL languages (Arabic, Hebrew):

```json
{
  "language": "ar",
  "ttsConfig": {
    "scriptDirection": "rtl"
  }
}
```

### Number Formatting

Configure how numbers are spoken:

```go
// Language-specific number formatting
type NumberFormat struct {
    Language        string
    GroupingSep     string  // "," or "."
    DecimalSep      string  // "." or ","
    SpokenFormat    string  // "one two three" vs "hundred twenty three"
}

var formats = map[string]NumberFormat{
    "en-US": {Language: "en-US", GroupingSep: ",", DecimalSep: ".", SpokenFormat: "natural"},
    "hi-IN": {Language: "hi-IN", GroupingSep: ",", DecimalSep: ".", SpokenFormat: "indian"},
    "de":    {Language: "de", GroupingSep: ".", DecimalSep: ",", SpokenFormat: "natural"},
}
```

### Honorifics

For languages with formal address:

```json
{
  "language": "hi-IN",
  "prompt": "Always use 'आप' (respectful you) instead of 'तुम'. Address customers with 'जी' suffix.",
  "llmConfig": {
    "formalityLevel": "formal"
  }
}
```

## Multi-lingual Agents

### Code-Switching Support

Handle users who switch languages mid-conversation:

```json
{
  "agent": {
    "name": "Bilingual Support",
    "primaryLanguage": "en-US",
    "secondaryLanguage": "hi-IN",
    "codeSwitchingEnabled": true,
    "prompt": "You can respond in both English and Hindi. Match the language the customer uses. Hinglish (mixed) is acceptable if the customer uses it."
  }
}
```

### Language Detection

```go
type LanguageDetector struct {
    stt STTProvider
}

func (d *LanguageDetector) Detect(audio []byte) (string, float32) {
    result := d.stt.DetectLanguage(audio)
    return result.Language, result.Confidence
}

// In pipeline
func (p *Pipeline) handleLanguageDetection(audio []byte) {
    lang, confidence := p.langDetector.Detect(audio)

    if confidence > 0.8 && lang != p.currentLanguage {
        p.switchLanguage(lang)
    }
}

func (p *Pipeline) switchLanguage(lang string) {
    // Update STT language
    p.stt.SetLanguage(lang)

    // Update TTS voice
    p.tts.SetVoice(getVoiceForLanguage(lang))

    // Update LLM context
    p.llm.SetLanguageContext(lang)
}
```

## Voice Selection by Language

### English Voices

| Provider | Voice | Style | Best For |
|----------|-------|-------|----------|
| Cartesia | Sophie | Warm, friendly | Customer support |
| ElevenLabs | Rachel | Professional | Sales |
| Azure | en-US-JennyNeural | Clear | IVR |

### Hindi Voices

| Provider | Voice | Style | Gender |
|----------|-------|-------|--------|
| Azure | hi-IN-SwaraNeural | Conversational | Female |
| Azure | hi-IN-MadhurNeural | Professional | Male |
| Google | hi-IN-Wavenet-A | Clear | Female |

### Spanish Voices

| Provider | Voice | Accent | Gender |
|----------|-------|--------|--------|
| Azure | es-MX-DaliaNeural | Mexican | Female |
| Azure | es-ES-ElviraNeural | Castilian | Female |
| ElevenLabs | Valentina | Neutral | Female |

## Accuracy Considerations

### Word Error Rate (WER) by Language

| Language | Deepgram | Google | Azure | ElevenLabs |
|----------|----------|--------|-------|------------|
| en-US | 8.4% | 10.2% | 11.5% | 12.1% |
| hi-IN | 22% | 12% | 15% | 18% |
| ta-IN | - | 14% | 18% | - |
| as-IN | - | 30% | - | 20% |

### Improving Accuracy

1. **Custom Vocabulary**: Add domain-specific terms

```json
{
  "sttConfig": {
    "keywords": ["Edesy:2", "voice agent:2"]
  }
}
```

2. **Prompt Engineering**: Help LLM understand context

```
"If the user's speech is unclear, ask them to repeat.
Common misheard words in Hindi: 'हां' (yes) and 'नहीं' (no)"
```

3. **Post-Processing**: Correct common errors

```go
func correctTranscript(text, language string) string {
    corrections := languageCorrections[language]
    for wrong, right := range corrections {
        text = strings.ReplaceAll(text, wrong, right)
    }
    return text
}
```

## Next Steps

- [Hindi Configuration](https://edesy.in/docs/voice-agent/languages/hindi) - Detailed Hindi setup
- [STT Providers](https://edesy.in/docs/voice-agent/stt) - Provider comparison
- [TTS Providers](https://edesy.in/docs/voice-agent/tts) - Voice selection