# Phantom-video/Phantom

Phantom: Subject-Consistent Video Generation via Cross-Modal Alignment

Repository: https://github.com/Phantom-video/Phantom
Canonical: https://ross.abutalabs.com/products/phantom
Homepage: https://phantom-video.github.io/Phantom/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: aigc, consistency-models, text-to-video, video-generation
Last push: 2025-09-11T14:52:42+00:00

## Health v2 (maintenance only)
Score: 39/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 41, release rhythm 35, longevity 40
- inputs: {"age_days": 569, "days_push": 356, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1517, forks 98 (observed 2026-08-28T04:04:57.315529+00:00)

## What it is
Phantom is a subject-consistent video generation model from ByteDance that preserves reference subject identity via cross-modal alignment. It includes inference code and checkpoints (including Phantom-Wan-14B) for subject-to-video generation.

## Use cases
- generate videos that preserve a reference subject's identity
- text-to-video generation with a reference image
- face identity preserving video generation
- subject-to-video generation research
- create consistent character videos from prompts

## When to choose
- you need videos that keep a reference face or subject consistent
- you want open weights for subject-driven video generation research
- you use ComfyUI workflows for video generation

## When to avoid
- you need lightweight or CPU-only video generation
- you want a polished end-user app rather than model code
- you need general video editing rather than generation

## Facets
- artifact type: library
- maturity: active
- function: video-processing, machine-learning, deep-learning
- domain: deep-learning, artificial-intelligence
- platform: python
- tags: text-to-video, subject-consistency, diffusion-models, aigc, identity-preservation, video, gpu

## Member repositories
- Phantom-video/Phantom (main) score 39

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:57.315529+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 2026-08-30T04:32:03.770400+00:00, confidence not recorded.
  - readme: https://github.com/Phantom-video/Phantom (fetched 2026-08-28T04:04:57.315529+00:00, sha 0d32a9532270)
  - homepage: https://phantom-video.github.io/Phantom/ (fetched 2026-08-29T11:35:26.546156+00:00, sha 5c373de73338)
- Data as of 2026-08-30T08:39:29.467469+00:00.