SANA-Streaming vs JoyAI-Video-Edit: inputs, controls and API cost
Compare SANA-Streaming and JoyAI-Video-Edit APIs: inputs, outputs, streaming delivery and ready-second cost. Choose a Windflow model for your product.
SANA-Streaming and JoyAI-Video-Edit are both live video transformation options in Windflow's public catalog. Choose by the inputs and interaction you need, then evaluate quality and responsiveness with your actual application. This page compares the published serving profiles; it does not present a head-to-head performance experiment.
Quick decision
- Evaluate SANA-Streaming for: Live restyling of a webcam feed: costume, setting or art-style changes while someone talks.
- Evaluate JoyAI-Video-Edit for: Instruction-based edits to live video: change clothing, background or style with a sentence.
Specifications side by side
| Published specification | SANA-Streaming | JoyAI-Video-Edit |
|---|---|---|
| Architecture | 2B streaming editing Diffusion Transformer | 16B multimodal DiT, chunk-autoregressive, two-step distilled |
| Resolution | 1280 × 704 native edit and delivery | 840 × 480 interactive tier; 1248 × 720 detail tier |
| Chunk | 24 frames | 8 frames |
| Control latency | A new edit prompt lands at the next chunk boundary | About 0.9 s from camera frame to the worker socket at 840 × 480, before the browser cushion |
| Inputs | A source clip you upload (≥ 41 frames) or a live webcam track, plus an edit prompt | A source clip you upload or a live webcam track, an edit instruction, and optionally a reference image |
| Outputs | The edited stream on main_video, with original motion preserved | The edited stream on main_video, with source motion and identity preserved |
| Delivery | WebSocket (H.264) · WebRTC | WebSocket (H.264) · WebRTC |
| Session protocol | sdk.realtime.v2 | sdk.realtime.v2 |
| Price | $0.0019 per ready second | $0.0021 per ready second |
| Implementation status | production | production |
Implementation status describes runtime maturity. Live GPU readiness is reported separately; Windflow is in private beta and brings capacity up on demand.
Input and control fit
Both models accept text prompts in the published Windflow profile. JoyAI-Video-Edit accepts reference or seed images; SANA-Streaming does not in its published profile. Both models accept uploaded source clips in the published Windflow profile. Both models accept live webcam input in the published Windflow profile. Both models do not accept movement and look controls in the published Windflow profile.
A shared session API does not make commands interchangeable. Use each model's supported inputs and control timing. In particular, prompt changes may land at a chunk or generation-window boundary; movement input can represent persistent state until replaced.
What SANA-Streaming is designed for
SANA-Streaming is a 2-billion-parameter streaming video-editing diffusion transformer from NVIDIA's SANA family. On Windflow you give it a live webcam track or an uploaded clip plus an edit prompt, and it returns the edited stream with the original motion preserved.
- Live restyling of a webcam feed: costume, setting or art-style changes while someone talks.
- Editing an uploaded clip from a text instruction without re-rendering from scratch.
- Interactive effects for streamers and video-call products.
- Prototyping prompt-driven video filters.
What JoyAI-Video-Edit is designed for
JoyAI-Video-Edit is a 16-billion-parameter autoregressive diffusion model from JD's open-source group for real-time, open-ended video editing. On Windflow you stream a webcam track or upload a clip, give an edit instruction (optionally with a reference image), and receive the edited stream with source motion and identity preserved.
- Instruction-based edits to live video: change clothing, background or style with a sentence.
- Reference-guided edits that carry a specific look onto a person or object.
- Higher-detail edits of uploaded clips using the 720p detail tier.
- Evaluating open-ended video editing before building a pipeline around it.
Compare a realistic session budget
- SANA-Streaming: $0.0019 per ready second, or $1.14 for ten ready minutes.
- JoyAI-Video-Edit: $0.0021 per ready second, or $1.26 for ten ready minutes.
Queueing and worker startup are free. A ready owner session can remain billable while delivery is paused or unwatched; release it when done. Independent sessions carry independent meters. Use the cost calculator to estimate your duration and daily sessions.
Evaluate before choosing
Use one repeatable input sequence, the same client and network, and equivalent output expectations. Record first visible frame, control-to-visible delay, frame gaps, scene continuity and rejected controls. Run several starts and a sustained session. If serving profiles differ, label the differences instead of treating displayed FPS as directly comparable native generation speed.
See the latency measurement guide and the model selector for a reproducible decision process.
Sources
- NVlabs/Sana on GitHub — Apache-2.0
- jd-opensource/JoyAI-Video-Edit on GitHub — Apache-2.0
FAQ
Is SANA-Streaming better than JoyAI-Video-Edit?
The answer depends on the inputs, control behavior, quality and output your product requires. The published specifications identify capability differences; an application-level evaluation establishes performance for your scenario.
Can I switch between these models through the same API?
Windflow keeps the session lifecycle common. Your application must adapt to each model's supported inputs, commands, output tracks and generation timing.
Do the published prices guarantee available GPU capacity?
No. Rates and implementation status are catalog facts. Live readiness is a separate operational signal, and private-beta access is required to run sessions.