Generated instantly from a single image.
Real-time lip-sync and interactivity for
the ultra-lightweight 2D mascot engine of the AI Avatar era.
Conventional 2D animation production required many steps for setup and tuning.
With MotionPortrait, just drop in an image.
No rigging — it starts moving as is.
* The comparison reflects general tendencies of other 2D animation engines, based on publicly available information from each company as of April 2026.
MotionPortrait is a dedicated engine that works across IP, apps, and devices.
One SDK that answers the challenges each use case faces.
FOR / 01
Publishers, game companies, animation studios, character businesses
Turn a dormant character into "someone who can talk."
Open up new touchpoints with fans through apps and streams.
FOR / 02
Startups, SaaS businesses, games, education services
An SDK that adds "facial expression" to LLM apps.
Can also be embedded into your existing AI stack via SDK.
FOR / 03
TVs, smart speakers, in-vehicle systems
Fully offline operation. An ultra-lightweight engine that brings "AI with facial expressions" to devices.
FOR / 04
Brands, agencies, digital signage
A next-generation ad experience where the character speaks to you.
Add a "face with expression" to banner and video ads.
Pick a "voice" below. All 6 avatars will
lip-sync at that voice's tempo.
Then select the avatar you want to speak.
From consumer games to corporate characters.
MotionPortrait is used across a wide range of fields.
Idea Factory / Compile Heart
Added lip-sync and facial expression changes to the 2D character art
of the long-loved goddess series. Strengthened the sense
of conversation with characters, raising the experience value across the series.
KADOKAWA CORPORATION
Adopted for the national IP "Haruhi Suzumiya" series app. Generated thousands of facial expression patterns in real time,
achieving both lip-sync and emotional expression.
Media Active Co., Ltd.
Adopted for a parenting app with 26 million cumulative downloads across the series.
The character realistically lip-syncs and moves
in time with audio, achieving a presence that
captivates children.
Lightweight, real-time, offline. Built for real-world deployment from the ground up.
Smooth performance with no loading wait.
Analyzes vowels/consonants to auto-generate mouth movement. Combine with ailia AI Speech for flawless Japanese too.
Compose expressions in real time with 7-axis sliders for joy, surprise, confusion, and more. Integrates naturally with LLM output.
One asset, fully supported on Unity, web browsers, iOS, Android, Windows, macOS, and Linux
across the board.
No cloud required. Works without internet access. Suited for offline environments
like healthcare, education, and customer service.
Set up from a single image. Fine-tune in the
dedicated editor when needed.
MotionPortrait runs on just text and voice input.
Connects directly to
your existing LLM/TTS.Integrates even more seamlessly with
ailia LLM / ailia AI Voice / AI Speech.
ailia SDK makes it even easier and smoother.
External API connections can be implemented in just a few lines too.
One unified API across Web, Unity, and native.
Just a few lines of code and your avatar starts moving.
// Animate an avatar on the web // Create and initialize the instance mpFace_ = new Module.MpFace(); mpRender_ = new Module.MpRender(); mpRender_.Init(); mpRender_.SetFace(mpFace_); // Load a pre-created avatar mpFace_.Load('avatar.bin'); // Just draw the avatar at a regular interval mpRender_.Draw();
We've gathered the questions we hear most often before adoption.
Feel free to contact us about anything else.
With the Business plan, you can embed it in games, apps, streams, ads, exhibits, and more.
Preset avatars are also OK for commercial use.
Yes. The engine itself runs entirely locally, with no cloud communication required at all.
It can be used on healthcare, caregiving, and in-store devices handling personal data without sending any data externally.
Yes. MotionPortrait works as long as it receives audio (WAV/PCM) and text. It can be used with any combination you like, including OpenAI, Anthropic Claude, ElevenLabs, VOICEVOX, Azure TTS, and more.
Just upload one image of your character to generate an avatar.
No parameter design or physics-simulation tuning like other technologies require.
If you need finer adjustments to expression or movement, you can also fine-tune it individually
in the dedicated editor.
If you're unsure about quality, our custom design service, described below, can also help.
WebGL2-capable browsers / Unity / iOS / Android / Windows / macOS / Linux
(major distributions)
are supported.
Runs at 60fps even on smartphones from a few generations ago.
It runs comfortably even on resource-constrained environments
such as home game consoles, embedded devices, and digital signage.
Yes. It supports the major ARM/x86 architectures.
Embedded licenses are also available for home appliances, signage, in-vehicle systems, smart speakers, and more,
and we also accept consultations for OEM/ODM mass production. For details, please contact us.
For a simple web integration, as fast as same day to a few days.
Even with full Unity integration and conversion of your own IP, adoption is possible
in 2-3 weeks.
For those who find presets not quite enough. We offer a custom design service.
Copyright © ailia Inc. / Axell Co., Ltd.
Contact Us
No rigging needed
