Real-time streaming avatar generator from a prompt and transcript, with motion control and consistent videos over 5 minutes. Runs in real time on a single H100; only a paper released.
Alibaba research model for real-time, long-horizon streaming character audio-video generation.
Updated EU status checked Last shown
StreamChar at a glance
- What is StreamChar?
- StreamChar is a model from Alibaba. Alibaba research model for real-time, long-horizon streaming character audio-video generation.
- Who makes StreamChar?
- StreamChar is made by Alibaba.
- Does StreamChar work in the EU?
- Unverified. No explicit evidence found. Only a research project page and paper were reported, with no public product or weights. Last verified 14 Nov 2026.
- How much does StreamChar cost?
- Pricing unknown. No public product or pricing found; research project page only.
- Where was StreamChar shown?
- In 1 creator video by AI Search. Most recently by AI Search on 7 Jun 2026, at 37:44 in “Full body waifus, AI dreams, realtime AI music, open-source Gemini Omni: AI NEWS”.
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
StreamChar is a research project from Alibaba's HumanAIGC group on long-horizon streaming character audio-video generation. According to a creator summary, it generates a streaming avatar from a prompt and transcript, with motion control and consistent output over more than five minutes.
The same source reports that it runs in real time on a single H100 GPU. The project page title and subtitle are the only details visible in the scraped material; the abstract was not captured, so these claims are not independently verified here.
At the time of the source material only a paper and project page were reported, with no released code, weights or public product. It is therefore of interest mainly to researchers and developers following real-time avatar and virtual-streamer generation.
- Streaming character audio-video generation
- Long-horizon consistency reported beyond 5 minutes
- Prompt and transcript conditioning
- Motion control
- Reported real-time inference on a single H100
As seen on YouTube
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Similar AI tools
Closest neighbours by meaning, not just by category.
Wan Streamer v0.3
Model · Alibaba
Alibaba's real-time audio-visual interactive character video model that renders speech and free-form actions.
Wan Streamer
Model · Alibaba
Alibaba research model for real-time conversational avatars using one transformer over text, audio and video.
Wan3.0
Model · Alibaba
Alibaba's Wan video generation model, reported to produce clips up to about 30 seconds, available via the Wan platform.
Happy Horse 1.1
Model · Alibaba
Alibaba's AI video generation model with native audio and lip sync, text, image and reference-to-video up to 1080p.
FashionChameleon
Model · Alibaba
Research model for real-time video virtual try-on that swaps garments mid-generation on a single GPU.
Wan Animate 2
Model · Alibaba
Alibaba's diffusion-transformer model that animates a character image from a driving video, with a real-time Lite variant.