Add a face to your Speechmatics pipeline

LemonSlice puts a lifelike, interactive avatar on top of the Speechmatics speech stack. Whether you run multilingual speech-to-text, Flow voice agents, or both, the interactive avatar turns that audio into natural, face-to-face conversation.

LemonSlice

At a glance

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

How the integration works

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

  1. Keep your Speechmatics setup

    Your Speechmatics transcription, language coverage, and Flow voice agent configuration stay untouched. LemonSlice sits downstream and only consumes the audio your pipeline already produces.

  2. Stream the audio to LemonSlice

    Send your agent's output audio to the LemonSlice avatar API. The avatar renders in real time with synchronized lip movement, facial expression, and gesture.

  3. The avatar video streams to your users

    LemonSlice delivers the rendered avatar video over WebRTC, natively via LiveKit, Daily (Pipecat), or Agora, so it drops into your existing client with no new infrastructure.

Drop in your real-time avatar with a short snippet

A Speechmatics 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 Speechmatics STT, plus your LLM and TTS, 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 interactive avatar that feels human

Control actions and emotions

Face, hands, shoulders, and posture all move through the Action Engine. 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 only interactive avatar model that handles non-humans, a turtle or a smiling teacup, with the same ease as cartoons and photorealistic people.

Ultra-low latency

LemonSlice delivers the fastest response time of any major avatar provider, so replies feel instant.

Create an avatar from one image

Any photo becomes an 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 Speechmatics run at production scale.

Thousands of concurrent calls

Run avatar sessions by the thousands at once, ready for high-volume production workloads.

Multi-region servers

Serve avatar sessions from the region nearest your users so latency stays minimal.

Zero data retention mode

Run sessions with zero stored conversation data, made for privacy-sensitive enterprise workflows.

What builders ship with Speechmatics + LemonSlice

Create avatars for everything from language practice to customer support, lead qualification, and corporate training.

Language practice partner

Practice real conversations in another language.

Customer support agent

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

Lead qualification agent

Qualify inbound leads in a face-to-face website conversation.

Frequently asked questions

Keep your Speechmatics setup as it is and stream your agent's output audio to the LemonSlice avatar API. The interactive avatar renders in real time and the video streams back to your users over WebRTC.
No. Speechmatics handles speech-to-text and the conversation through Flow, while LemonSlice handles the visual layer, a real-time avatar that speaks whatever your pipeline produces.
Yes. LemonSlice animates the audio your pipeline sends, so the language and accent coverage you get from Speechmatics carries through to the avatar unchanged.
Yes. Flow produces the conversational audio, that audio streams to the LemonSlice avatar API, and the rendered avatar video streams to your users in real time.
Any Speechmatics pipeline that produces output audio works, from Flow voice agents to custom stacks built on Speechmatics transcription. LemonSlice only needs the audio stream your agent already generates.
No. LemonSlice responds faster than any other major avatar provider, and multi-region servers keep sessions close to your users, so every conversation keeps feeling instant.

Give your Speechmatics agent a face

Keep your pipeline, add a real-time avatar with a 5-minute API setup.