Published by Video Runtime · Source check 2026-09-21 · Evidence method · Suggest a correction
EDITORIALLY VERIFIED PRODUCT INFORMATION
About JoyAI-Echo
Two released research systems in one repository: a long-horizon multi-shot audio-video generator and an interactive omnimodal world model. Echo-WM is an omnimodal world model that responds to continuous navigation while video, environmental sound, music and speech evolve together in one framework. Echo-LongVideo carries characters and scenes across shots with a paired audio-video memory bank for long-horizon multi-shot generation.
Key features
Ships both an interactive world model and a ten-minute-plus long-video system with inference code, Gradio demos and Hugging Face checkpoints. The surrounding ecosystem includes a ComfyUI integration and a released Unreal Engine simulation pipeline for the Echo-WM world data engine.
Getting started
Pick Echo-LongVideo or Echo-WM, create the documented conda environment for that subproject, download its listed checkpoints from Hugging Face, then run the supplied inference entrypoint or Gradio demo before customising prompts.
Real-time interaction
Echo-WM is an omnimodal world model that responds to continuous navigation while video, environmental sound, music and speech evolve together in one framework. Echo-LongVideo carries characters and scenes across shots with a paired audio-video memory bank for long-horizon multi-shot generation.
Pricing
LTX-2 Community License code; self-hosted compute costs extra Repository code is published under LTX-2 Community License. Confirm separate model, dataset and dependency terms before reuse. No hosted price is inferred. Budget for the documented GPU, storage and setup requirements.
Limitations
The repository uses the LTX-2 Community License rather than a standard permissive open-source licence. The two projects keep separate Python environments and separately downloaded checkpoint directories.
EDITORIAL VIEW
Editor's Verdict
JoyAI-Echo is included as a source-available implementation relevant to real-time generative video or interactive world systems.
Run the supplied example on representative hardware and review every linked licence before production use.
How It Works
- Echo-WM is an omnimodal world model that responds to continuous navigation while video, environmental sound, music and speech evolve together in one framework.
- Echo-LongVideo carries characters and scenes across shots with a paired audio-video memory bank for long-horizon multi-shot generation.
What We Like
- Ships both an interactive world model and a ten-minute-plus long-video system with inference code, Gradio demos and Hugging Face checkpoints.
- The surrounding ecosystem includes a ComfyUI integration and a released Unreal Engine simulation pipeline for the Echo-WM world data engine.
Current Limitations
- The repository uses the LTX-2 Community License rather than a standard permissive open-source licence.
- The two projects keep separate Python environments and separately downloaded checkpoint directories.
Pricing & Access
- Repository code is published under LTX-2 Community License. Confirm separate model, dataset and dependency terms before reuse.
- No hosted price is inferred. Budget for the documented GPU, storage and setup requirements.
Verification Summary
- Official repository and licence reviewed on 2026-09-21. Video Runtime has not installed or benchmarked this project.
- Performance and compatibility statements below are attributed to the maintainers and are not site measurements.
Technical notes & official performance
Official Performance Data
- The README states Echo-LongVideo supports 10+ minutes of long-horizon generation with continuity across shots. This is a maintainer claim; Video Runtime has not reproduced it.
Technical Notes
- The 1.5 release (2026-08-28) added long-horizon generation, consumer-GPU inference profiles and a Director Agent; the Echo-WM world model and an Unreal Engine data-engine pipeline followed in separate releases.
- Research papers cover JoyAI-Echo 1.5, the Echo-WM world model and the UE pipeline on arXiv.
Sources
- Official GitHub repository ↗github · accessed 2026-09-21
- Code licence ↗github · accessed 2026-09-21
- Official project page ↗official-docs · accessed 2026-09-21
Access, stage & mechanism
These are separate source-checked facts, not a live uptime monitor. Unverified fields are left open; transport and persistent sessions do not establish how frames are generated.
- Release stage
- Not verified
- Access mode
- Not verified
- Source availability
- Not verified
- Transport
- Not verified
- Interaction
- Not verified
- Session mode
- Not verified
- Generation mechanism
- Not verified
- Continuity method
- Not verified
- Licence status
- Not verified
- Tasks
- Not classified
CONTINUE EXPLORING
Related projects & models
Selected by research links and category, not by popularity.
Build long-horizon playable video worlds with camera control, prompt switching and explicit spatial memory.
Another frame streaming option in this section.Public repository; self-hosted setup and compatible GPU requiredReal-time instruction-guided editing of open-ended video streams with an autoregressive diffusion transformer.
A related project selected in this entry’s research.Public repository; self-hosted installation and compatible GPU required
Run an action-conditioned interactive world model with open-ended rollout on a single desktop GPU.
Another frame streaming option in this section.Public repository; self-hosted installation and compatible GPU requiredGenerate continuing interactive worlds with bounded external state, mid-rollout prompts and camera control.
Another frame streaming option in this section.Public repository; self-hosted setup and compatible GPU requiredRead how official-source and hands-on records stay separate in our Methodology.