Simli

Overview

Simli is a speech to video API that adds a lip-synced video face to a real-time AI agent, streamed over WebRTC at under 300 milliseconds of added latency. Its Trinity-1 Gaussian model renders facial expression and head movement live rather than pre-rendering a clip. Developers pick from 21 documented default faces or upload a face image in the web app to build their own, then drive it from JavaScript and Python SDKs or through open-source plugins for LiveKit Agents and Pipecat. Billing is usage-based, with $10 of credit on signup and a monthly top-up of 50 minutes on the free plan.

Best for developers adding a real-time video face to a voice AI agent they already run.

Simli is a component, not a product, and that is the whole decision: it renders the face and leaves speech recognition, the language model and the voice to whatever stack you already run. That suits a team building an interactive agent into its own app, where the avatar has to react inside a conversation rather than read a finished script. Anyone who wants a finished video file will find nothing here to click, because there is no editor and no render queue, only a stream. The per-minute meter rewards short sessions and punishes long ones, so the cost of a feature scales with how much people actually talk to it, which is worth modelling before the launch rather than after.

Key Facts

Features

Avatar library

Yes, 21 default faces documented for immediate use

Custom and cloned avatars

Yes, upload a face image in the web app; processing takes a couple of hours

Languages and translation

No list of its own. Simli lip-syncs whatever audio it is handed, so the spoken languages come from the speech services in your own stack; for Simli Auto the documentation points at the language list of its speech-to-text provider, Deepgram, and names ElevenLabs as the text-to-speech example

Editor

No, there is no video editor: faces are created in the web app and driven from your own application code

Learning features

No, Simli is an avatar streaming layer without courses, quizzes or SCORM export

API

Yes, a speech to video API with JavaScript and Python SDKs, plus open-source plugins for LiveKit Agents and Pipecat

Integrations

• LiveKit
• Pipecat

Deployment and Support

Deployment

• Cloud (SaaS)
• Browser-based

Support

• Community forum

Onboarding

• Documentation and knowledge base

Pros and Cons

Pricing

• Free

  $0

  $10 of credit on signup and a monthly top-up of 50 minutes.

• Pay as you go

  Not stated

  Usage-based billing with volume discounts.

Best For

• Technology

Alternatives and Comparisons

What users say

Reviews

Share your experience

Write a Review