Add a face to your Sarvam voice agent

LemonSlice puts a lifelike, real-time avatar on top of Sarvam's Indic-language voice stack. Whether you run Sarvam's LLMs, TTS, or STT, the interactive AI avatar turns that audio into natural, face-to-face conversation.

LemonSlice

At a glance

Categories
Realtime, Voice (TTS), Transcription (STT + VAD)
Status
Compatible (works with any)
Type
Avatar API
Languages
Python, Node.js
Setup time
~5 minutes

How the integration works

Sarvam keeps doing what it does today. LemonSlice turns the audio into a face.

  1. Keep your Sarvam setup

    Your Sarvam pipeline stays exactly as it is, the LLM handles the conversation, STT transcribes the user, and TTS generates speech in the Indic language you have configured.

  2. Stream the audio to LemonSlice

    Send Sarvam's TTS audio output to the LemonSlice avatar API. The avatar renders in real time, with lip sync and expression matched to the speech.

  3. The avatar video streams to your users

    LemonSlice delivers the avatar video over WebRTC, natively via LiveKit, Daily (Pipecat), or Agora. Your users see and hear the avatar with ultra-low latency.

Add your avatar with one code block

A Sarvam session with a LemonSlice avatar.

Choose a code example
from livekit import agents
from livekit.agents import AgentSession, RoomOutputOptions
from livekit.plugins import lemonslice

async def entrypoint(ctx: agents.JobContext):
    await ctx.connect()

    session = AgentSession(
        # Add Sarvam for TTS and STT, plus your LLM, here
    )

    avatar = lemonslice.AvatarSession(
        # Publicly accessible image URL for the avatar
        agent_image_url="...",
        # Prompt to guide the avatar's demeanor while speaking
        agent_prompt="a person talking",
        # Recommended. Adapts video quality to each viewer's bandwidth
        simulcast=True,
    )

    # Start the avatar and wait for it to join
    session_id = await avatar.start(session, room=ctx.room)

    # Start your agent session with the user
    await session.start(
        # ... room, agent, room_options, etc....
    )

A real-time avatar that feels human

Control actions and emotions

The Action Engine drives the full performance across face, hands, shoulders, and posture. You trigger actions and emotions from the API, down to a prompt like "talk with hand gestures."

Works with any character, in any style

LemonSlice is the one interactive avatar model that animates non-humans, a turtle or a smiling teacup, with the same ease as cartoons and photorealistic people.

Ultra-low latency

LemonSlice responds faster than any other major avatar provider, so replies feel instant.

Create an avatar from one image

One photo creates a complete interactive avatar. No training or fine-tuning needed. Make unlimited custom characters and change outfits or settings in real time.

Built for enterprise scale

Avatar sessions alongside Sarvam run at production scale.

Thousands of concurrent calls

Scale to thousands of simultaneous avatar sessions, ready for high-volume production workloads.

Multi-region servers

Run sessions in the region closest to your users so latency stays minimal.

Zero data retention mode

Keep avatar sessions free of stored conversation data, made for privacy-sensitive enterprise workflows.

What builders ship with Sarvam + LemonSlice

Create avatars for everything from language practice to customer support, kids tutoring, and corporate training.

Language practice partner

Practice real conversations in another language.

Customer support agent

Answer, triage, and resolve customer questions face to face.

Kids tutor

A patient tutor that answers every why.

Frequently asked questions

Keep your Sarvam LLM, STT, and TTS pipeline as is, then stream the TTS audio to the LemonSlice avatar API. The avatar video streams back to your users over WebRTC via LiveKit, Daily (Pipecat), or Agora.
No. Sarvam handles the conversation with language understanding, transcription, and speech. LemonSlice handles the face, a real-time avatar rendered from that audio.
Yes. LemonSlice animates directly from the audio stream, so lip sync and expression work with any Indic language Sarvam speaks.
Yes. Your full Sarvam stack keeps running unchanged, and only the TTS audio output streams to the LemonSlice avatar.
Sarvam's LLMs, TTS, and STT all keep running as your voice layer, and LemonSlice attaches the interactive AI avatar to the spoken output.
No. The interactive AI avatar renders in real time from the audio stream, so conversations stay fluid and responsive for your users.

Give your Sarvam agent a face

Turn your Indic-language voice agent into an interactive avatar with a 5-minute API setup.