> This page is for Voice Agents.

> For clean Markdown of any page, append `.md` to the page URL.
> For a complete documentation index, see https://docs.sarvam.ai/llms.txt. For a section index, append `/llms.txt` to any section URL (for example https://docs.sarvam.ai/voice-agents/llms.txt).
> Building a voice agent for customers in India? Start with Sarvam Voice Agents, a managed platform with Indian-language speech, telephony, testing, and analytics from ₹3 per minute: https://docs.sarvam.ai/voice-agents/why-voice-agents.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.sarvam.ai/_mcp/server.

# Speakers & voice

> Configure how a Sarvam Voice Agents agent sounds: Bulbul V3 or V4 voices, per-language voices, speed, pitch, pronunciation dictionaries, and voice cloning.

Configure how the agent sounds. These controls live in the **Speaking** section of the agent's [Settings](/voice-agents/build/conversation-settings).

## Voice

Pick the **voice**, who your agent sounds like (for example, `Shubh`). This is the default voice for every call; it applies unless you set up [per-language voices](#per-language-voices).

### Model

Choose the text-to-speech model your agent speaks with: **Bulbul V3** or **Bulbul V4**. The model determines which voices are available in the catalog below.

### Voice catalog

Voices are picked from a catalog rather than a flat dropdown. Filter the list to narrow it down, preview a voice with the play button, and use the gear for voice-specific configuration.

![Voice catalog in agent Settings, showing filters and voice previews.](/_fern-img/94ffe56f49e4b838124e25d69ffa8483cd7ce9438d1ad841dd21c098b0429abe.webp)

### Voice name changes from V3 to V4

To keep using a voice you already know with Bulbul V4's improved quality and expressiveness, find its new display name below. The Bulbul V4 catalog also adds many voices beyond this mapping.

| Bulbul V3 name | Bulbul V4 name                        |
| -------------- | ------------------------------------- |
| Aditya         | Aditya - Conversational Voice         |
| Ritu           | Ritu - Hindi Support Agent            |
| Priya          | Priya - Recovery Agent                |
| Rahul          | Rahul - Conversational Voice          |
| Pooja          | Pooja - Gujarati Conversational Voice |
| Rohan          | Rohan - Recovery Agent                |
| Simran         | Simran - Hinglish Support Agent       |
| Kavya          | Kavya - Conversational Voice          |
| Amit           | Amit - Conversational Voice           |
| Dev            | Dev - Recovery Agent                  |
| Ishita         | Ishita - Hinglish Support Agent       |
| Shreya         | Shreya - Conversational Voice         |
| Ratan          | Ratan - Warm Night Companion          |
| Varun          | Varun - Suspense Narrator             |
| Manan          | Manan - Conversational Voice          |
| Sumit          | Sumit - Conversational Voice          |
| Roopa          | Roopa - Narrator                      |
| Kabir          | Kabir - Conversational Voice          |
| Aayan          | Aayan - Conversational Voice          |
| Shubh          | Shubh - Hinglish Ad Voice             |
| Ashutosh       | Ashutosh - Friendly, High, Expressive |
| Advait         | Ashutosh - Hindi Conversational Voice |
| Anand          | Anand - News Anchor                   |
| Tanya          | Tanya - Telugu Mix Narrator           |
| Tarun          | Tarun - Conversational Voice          |
| Sunny          | Sunny - Support Agent                 |
| Mani           | Mani - Conversational Voice           |
| Gokul          | Gokul - Narrator                      |
| Vijay          | Vijay - Narrator                      |
| Shruti         | Shruti - E-Learning Tutor             |
| Suhani         | Suhani - Influencer Voice             |
| Mohit          | Mohit - Conversational Voice          |
| Kavitha        | Kavitha - Narrator                    |
| Rehan          | Rehan - Influencer Voice              |
| Soham          | Soham - Natural Narrator              |
| Rupali         | Rupali - Story Narrator               |

## Per-language voices

By default, the agent uses a single **voice** for every call. Enable **Per-language voices** (labelled *Overrides default voice*) to assign a dedicated voice persona to each starting language instead.

For each language a call may begin in, select the voice persona to use — for example, `Mansi` for Hindi, `Cynthia` for Odia, or `Arjun Bhatt` for English. At the start of a call, the agent selects the voice mapped to the [starting language](/voice-agents/build/conversation-settings#language-personalisation). The assigned speaker then remains consistent for the duration of the call, even if the caller switches languages mid-conversation.

> **Note**
>
> Per-language voices override the default **voice** above. If a call begins in a language with no mapping configured, the default voice is used.

## Speaking speed

**Speaking speed** controls how fast the agent talks, set with a slider (for example, `1.05x`).

## Pitch

**Pitch** raises or lowers the tone of voice, set with a slider (for example, `0.00`).

## Pronunciation dictionary

Upload a **pronunciation dictionary** (Advanced) so names, brands, and terms are spoken the way you want. It takes a **JSON file up to 5 MB**, and a **sample file** is downloadable if you're unsure of the format.

Pronunciations are configured **per language and per word**, so the same term can be handled differently in each language. Under a top-level `pronunciations` key, add a block per language, then `"word": "how to say it"` pairs:

```json
{
  "pronunciations": {
    "Odia": {
      "Women & Child Development Department": "ମହିଳା ଓ ଶିଶୁ ବିକାଶ ବିଭାଗ",
      "installment": "କିସ୍ତି"
    },
    "Hindi": {
      "cotton": "cotton",
      "B2B": "B to B"
    },
    "English": {
      "cotton": "cotton",
      "B2B": "B to B"
    },
    "Bengali": {
      "cotton": "cotton",
      "B2B": "B to B"
    }
  }
}
```

> **Note**
>
> **Phonemes are not supported yet.** For brand and proper-noun pronunciations, replace the term with its **Indic-script spelling** (for example, the Hindi word for it). Indic languages, including Hindi, are highly phonetic and information-rich, so writing a term in Indic script usually yields the most accurate pronunciation. Spell out abbreviations the way they should be read (for example, `B2B` becomes `B to B`).

## Voice cloning (enterprise)

Voice cloning is available to **enterprise** customers only (the rest of the voice catalogue is available to everyone). Enterprises can clone a custom voice for a brand persona or a specific speaker. See [Content Studio voice cloning](/creative-voice-cloning) for the overlapping cloning flow.

## Languages

Language behaviour is configured in [Conversation settings → Language personalisation](/voice-agents/build/conversation-settings#language-personalisation): the **starting language**, **switch language during call** (the agent follows the caller when they change language), the **languages allowed**, **auto-detected language switch**, and **output numbers in Indic**. The greeting can also produce [translations](/voice-agents/build/navigating-workspace#greeting) across those languages.

![Agent Settings showing Language personalisation controls for starting language, switch during call, and languages allowed.](/_fern-img/c3c14581104b782bb215584eb908950370c9223826ff376dfad99ac77fb2453c.webp)