Live video transformation
JoyAI-Video-Edit API
A source-guided edited stream on main_video. Served on Windflow as a realtime session you can plug into your product and pay for by the second.
Overview
What JoyAI-Video-Edit does
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 and an optional reference image, and receive an edited stream guided by those inputs. Source motion and identity are evaluation criteria, not guaranteed invariants.
A session keeps the model running: connect once, steer it with text prompts, a reference image, a source video and a live webcam feed while it runs, and receive the live stream back on a named output track.
Every model in the catalog speaks the same typed session protocol, so switching between models changes the controls you send, not how you integrate.
- Architecture
- 16B multimodal DiT, chunk-autoregressive, two-step distilled
- Resolution
- 840 × 480 interactive tier; 1248 × 720 detail tier
- Chunk
- 8 frames
- Control latency
- About 0.9 s from camera frame to the worker socket at 840 × 480, before the browser cushion
- Inputs
- A source clip you upload or a live webcam track, an edit instruction, and optionally a reference image
- Outputs
- A source-guided edited stream on main_video
- Delivery
- WebSocket (H.264) · WebRTC
- Session protocol
- sdk.realtime.v2
- Price
- $0.0021 per billable second
$0.0021
Per billable second
Meter starts when generation begins
$7.56
Per billable hour
The same rate, held for an hour
Production
Runtime status
Capacity is brought up on demand
Private beta
Access
Request access or a live demo
Use cases
What teams build with JoyAI-Video-Edit
- 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.
Provenance
Where JoyAI-Video-Edit comes from
The model and its code are published by the upstream authors. Windflow runs it behind one session API and bills per billable second.
- jd-opensource/JoyAI-Video-Edit on GitHubLicence · Apache-2.0
Compare
JoyAI-Video-Edit next to similar models
| Model | Inputs | Output | Price per billable second |
|---|---|---|---|
| JoyAI-Video-Edit | A source clip you upload or a live webcam track, an edit instruction, and optionally a reference image | A source-guided edited stream on main_video | $0.0021 |
| SANA-Streaming | A source clip you upload (≥ 41 frames) or a live webcam track, plus an edit prompt | A source-guided edited stream on main_video | $0.0019 |
| EdenLive | Source game footage or a live video track, a scene prompt, and an optional visual reference | A proxy-guided generated video stream on main_video | $0.0021 |
FAQ
Questions about JoyAI-Video-Edit
- What resolutions does JoyAI-Video-Edit offer?
- Two tiers: an interactive tier at 840 × 480 and a detail tier at 1248 × 720. Super-resolution can be applied to the delivered stream.
- Is an edit applied to the whole clip?
- The instruction describes the intended change, but output is generated video. Unmentioned details, motion or identity can change. Evaluate the complete clip and the transition after an instruction update, not only one selected frame.
- Does the published latency experiment measure my browser's response?
- No. The recorded MI355X experiment uses a worker-loopback probe and separates camera-frame-to-edited-pixels from edited-pixels-to-socket timing. Capture, network, browser decode and presentation need evaluation on your actual client path.
- Can I add the latency-stage medians to estimate a total median?
- No. An end-to-end median or tail percentile requires paired start/end timestamps for the same frames. Independent stage percentiles cannot establish the percentile of the complete path.
- How much does JoyAI-Video-Edit cost on Windflow?
- JoyAI-Video-Edit costs $0.0021 per billable session-second ($7.56/hr of billable time). Billing starts at the session's first generation_started event and stops when the session is released or ends. Queueing, worker startup, time before generation starts and the documented reconnect grace are not billed. After generation starts, paused or unwatched time remains billable; release the owner session to stop the meter. There is no subscription.
- How do I get access to JoyAI-Video-Edit?
- Windflow is in private beta. Request access from the sign-up page, or ask for a live demo of JoyAI-Video-Edit. Approved accounts start with $5 and four concurrent sessions.
- Is JoyAI-Video-Edit always running?
- No. Model capacity is brought up on demand rather than kept standing. Live readiness for each model is published at api.windflow.online/v1/status, and a live demo can be arranged on request.
Guides and comparisons
- GuideChoosing a realtime video model: a decision guide for 12 hosted models
- Evaluation guideHow to measure realtime video API latency and responsiveness
- Measured engineeringJoyAI live editing latency: what the 443 ms measurement includes
- Cost comparisonRealtime video API pricing: compare rates and billable timelines
- API guideRealtime video API: generate, edit and control live AI video
- Live editing comparisonWindflow vs Decart: evaluating a live video editing API
- Endpoint comparisonWindflow vs fal: choosing the right realtime video endpoint
- Model API guideHow to use the JoyAI Video Edit API
- Prompt guideHow to write edit instructions for realtime video editing
- Use caseLive video editing from a webcam with a text instruction
- Troubleshooting guideHow to recover a realtime video stream after a dropped connection
Other models