Z.aiRealVideo
RealVideo
A modular research system for conversational video.
Research release
Practical use
Practical guidance. Plan for the latency of the full chain: text, speech and video. Inspect lip alignment during a rapid change of sentence or emotion, where errors can appear even when a single short clip looks correct.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | realvideozai-org/RealVideo |
| Ancestry | Video component derives from Wan2.2-S2V-14B.RealVideo official documentation |
| Components | GLM-4.5-AirX and GLM-TTS generate responses; autoregressive diffusion produces video frames.RealVideo official documentation |
| Licence | MIT metadata.RealVideo official documentation |