# jd-opensource/JoyAI-Echo

JoyAI-Echo: Pushing the Frontier of Long Audio-Visual Generation

Repository: https://github.com/jd-opensource/JoyAI-Echo
Canonical: https://ross.abutalabs.com/products/joyai-echo
Homepage: https://echo-team-joy-future-academy-jd.github.io/Echo-LongVideo-Page/
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-08-26T17:50:55+00:00

## Health v2 (maintenance only)
Score: 58/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 35, longevity 6
- inputs: {"age_days": 92, "days_push": 7, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1943, forks 169 (observed 2026-08-28T04:05:56.427280+00:00)

## What it is
JoyAI-Echo is a Python framework for long-horizon audio-visual generation, producing coherent multi-shot videos up to ~5 minutes with paired audio-video memory banks, plus an omnimodal interactive world model (Echo-WM). It includes inference scripts, Gradio demos, and a ComfyUI integration, released for non-commercial research use.

## Use cases
- generate minute-long coherent videos with consistent characters and voice
- build an interactive generative world model that responds to navigation
- create multi-shot stories with synchronized audio and video
- run conversational editing of generated video in real time
- research long-video diffusion and memory-based consistency

## When to choose
- you need long-form (minutes) audio-visual generation with cross-shot consistency
- you want an omnimodal world model for interactive media research
- you work in Python with GPU inference and want ComfyUI support

## When to avoid
- you need a commercially licensed video generation model (non-commercial only)
- you need short-clip fast generation without long-horizon memory
- you lack GPU resources for large diffusion transformer inference

## Facets
- artifact type: library
- maturity: active
- function: video-processing, audio-processing, machine-learning, deep-learning, llm-inference
- domain: deep-learning, artificial-intelligence, computer-vision
- platform: python, cli
- tags: video-generation, world-model, audio-visual, long-video, diffusion-transformer, text-to-video, comfyui, non-commercial, video, audio, linux, gpu

## Member repositories
- jd-opensource/JoyAI-Echo (main) score 58

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:56.427280+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:08:01.946641+00:00, confidence not recorded.
  - readme: https://github.com/jd-opensource/JoyAI-Echo (fetched 2026-08-28T04:05:56.427280+00:00, sha a8e7304e24f0)
  - homepage: https://echo-team-joy-future-academy-jd.github.io/Echo-LongVideo-Page/ (fetched 2026-08-29T10:47:05.549713+00:00, sha 4dc31b68880f)
- Data as of 2026-08-30T08:39:29.467469+00:00.