# SkyworkAI/SkyReels-V1

SkyReels V1: The first and most advanced open-source human-centric video foundation model

Repository: https://github.com/SkyworkAI/SkyReels-V1
Canonical: https://ross.abutalabs.com/products/skyreels-v1
Homepage: https://www.skyreels.ai
Language: Python
License: NOASSERTION
License Family: other
Topics: i2v, t2v, video-diffusion-transformers
Last push: 2025-03-10T10:02:10+00:00

## Health v2 (maintenance only)
Score: 25/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 10, release rhythm 35, longevity 40
- inputs: {"age_days": 567, "days_push": 541, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2696, forks 311 (observed 2026-08-28T04:07:11.197266+00:00)

## What it is
SkyReels V1 is an open-source human-centric video foundation model with Text-to-Video and Image-to-Video variants, fine-tuned from HunyuanVideo on millions of high-quality film and television clips. The repository provides model checkpoints, inference code, a Gradio web demo, and multi-GPU parallel inference support including consumer-level RTX 4090 inference.

## Use cases
- generate cinematic videos from a text prompt
- animate a still portrait image into a video
- run video diffusion inference on a single RTX 4090
- generate human-centric footage with realistic facial expressions
- parallelize video generation across multiple GPUs
- compare open-source text-to-video models against Kling or Hailuo
- self-host a gradio demo for video generation

## When to choose
- you need open weights for text-to-video or image-to-video generation with a focus on human faces and cinematic quality
- you want to run video generation inference locally on consumer GPUs or scale it across multiple GPUs
- you are building on top of HunyuanVideo-style diffusion transformers and want a film-tuned checkpoint

## When to avoid
- you need 720p output, CFG-distilled or lite models, or ComfyUI integration - these are still on the TODO list
- you have no GPU available, since inference requires substantial VRAM
- you need a fully permissive license - the license is marked NOASSERTION, so verify terms before commercial use

## Facets
- artifact type: library
- maturity: active
- function: video-processing, machine-learning, deep-learning, llm-inference
- domain: artificial-intelligence, image-processing, deep-learning, media
- platform: python
- tags: text-to-video, image-to-video, diffusion-transformer, video-generation, human-centric-video, hunyuanvideo, generative-ai, gradio-demo, multi-gpu-inference, film-generation, video, gpu, linux, docker

## Member repositories
- SkyworkAI/SkyReels-V1 (main) score 25

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:11.197266+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 2026-08-30T02:16:18.730504+00:00, confidence not recorded.
  - readme: https://github.com/SkyworkAI/SkyReels-V1 (fetched 2026-08-28T04:07:11.197266+00:00, sha ada7f77d0b01)
  - homepage: https://www.skyreels.ai (fetched 2026-08-29T09:59:21.695069+00:00, sha f93c417b00dc)
- Data as of 2026-08-30T08:39:29.467469+00:00.