Add a face to your OpenAI voice stack

LemonSlice puts a lifelike, real-time avatar on top of OpenAI's voice stack. Whether you run the Realtime API, TTS voices, or Whisper transcription, the interactive avatar turns that audio into natural, face-to-face conversation with your customers.

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

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

  1. Keep your OpenAI setup

    Your Realtime API sessions, TTS voices, or Whisper transcription pipeline stay untouched. OpenAI keeps handling the conversation and the audio.

  2. Stream the audio to LemonSlice

    Send the audio output from your OpenAI pipeline to the LemonSlice avatar API. The avatar renders in real time with accurate lip sync and expression.

  3. The avatar video streams to your users

    LemonSlice delivers the avatar video over WebRTC, natively via LiveKit, Daily (Pipecat), or Agora, so it drops into your existing transport.

Add your real-time avatar with a few lines of code

A OpenAI Realtime 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 OpenAI 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....
    )

An avatar that listens, reacts, and feels human

Control actions and emotions

The Action Engine animates face, hands, shoulders, and posture. You control actions and emotions through the API, down to a prompt like "talk with hand gestures."

Works with any character, in any style

LemonSlice is the only interactive avatar model that animates non-humans (a smiling teacup, a turtle) as easily as photorealistic people and cartoons.

Ultra-low latency

LemonSlice has the fastest response time of any major avatar provider, keeping conversations feeling instant.

Create an avatar from one image

From one photo comes a full interactive avatar. No training or fine-tuning required. Create unlimited custom characters with real-time outfit or setting changes.

Built for enterprise scale

Avatar sessions alongside OpenAI Realtime run at production scale.

Thousands of concurrent calls

Run thousands of simultaneous avatar sessions, designed for high-volume production workloads.

Multi-region servers

Deploy avatar sessions from the region closest to your users to keep latency low.

Zero data retention mode

Run avatar sessions without storing conversation data, designed for privacy-sensitive enterprise workflows.

What builders ship with OpenAI + LemonSlice

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

Kids tutor

A patient tutor that answers every why.

Interviewer

Structured screening interviews, face to face.

Corporate training coach

Roleplay difficult conversations and train staff on policy.

Frequently asked questions

Keep your OpenAI pipeline as is and stream its audio output to the LemonSlice avatar API. The real-time avatar renders with accurate lip sync, and the video streams back over WebRTC via LiveKit, Daily (Pipecat), or Agora.
No. OpenAI handles the conversation, speech, or transcription, and LemonSlice handles the visual layer, a real-time avatar with lip sync and expression.
All three audio lanes pair with the avatar. The Realtime API for speech-to-speech, TTS voices, and Whisper-lineage transcription each produce audio that LemonSlice animates in real time.
Yes. LemonSlice animates the audio OpenAI produces, so the voice your users hear stays exactly the same.
No. LemonSlice delivers the fastest response time of any major avatar provider, so conversations still feel instant with the avatar layer in place.
Yes. LemonSlice offers a zero data retention mode built for privacy-sensitive workflows, and multi-region servers keep sessions close to your users.

Give your OpenAI agent a face

Keep your pipeline, add a face with a 5-minute API setup.