# KlingAIResearch/ReCamMaster

[ICCV'25 Best Paper Finalist] ReCamMaster: Camera-Controlled Generative Rendering from A Single Video

Repository: https://github.com/KlingAIResearch/ReCamMaster
Canonical: https://ross.abutalabs.com/products/recammaster
Homepage: https://jianhongbai.github.io/ReCamMaster/
Language: Python
License: MIT
License Family: permissive
Topics: aigc, computer-vision, video-generation, 4d, camera-control
Last push: 2025-11-28T04:06:03+00:00

## Health v2 (maintenance only)
Score: 44/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 54, release rhythm 35, longevity 38
- inputs: {"age_days": 537, "days_push": 278, "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 1855, forks 98 (observed 2026-08-28T04:05:44.805346+00:00)

## What it is
ReCamMaster is a reference implementation of a camera-controlled generative video rendering model that re-renders a single source video along novel camera trajectories using a diffusion-based video model. It includes training and inference code, a model checkpoint built on Wan2.1, and the MultiCamVideo multi-camera dataset rendered in Unreal Engine 5.

## Use cases
- re-render a video with a new camera trajectory
- generate novel camera angles from a single video
- stabilize shaky handheld footage by applying smooth camera paths
- generate multi-view videos for 4D reconstruction
- create cinematic camera moves like pans, zooms, and arcs on existing footage
- train a camera-controllable video generation model

## When to choose
- you need to change or synthesize camera motion for an existing video
- you want a research-grade codebase and dataset for camera-controlled video generation
- you need multi-view video data for 4D reconstruction or embodied AI research

## When to avoid
- you need production-quality output matching the paper's demos - the open-source version lags the hosted model
- you want a simple end-user app rather than running Python inference on a GPU
- you need real-time video processing

## Facets
- artifact type: library
- maturity: active
- function: video-processing, machine-learning, deep-learning, computer-vision
- domain: computer-vision, artificial-intelligence, deep-learning
- platform: python
- tags: video-generation, camera-control, diffusion-model, 4d-reconstruction, video-stabilization, research-code, iccv-2025, multicamvideo-dataset, video, gpu, linux

## Member repositories
- KlingAIResearch/ReCamMaster (main) score 44

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:44.805346+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 2026-08-30T03:16:43.675855+00:00, confidence not recorded.
  - readme: https://github.com/KlingAIResearch/ReCamMaster (fetched 2026-08-28T04:05:44.805346+00:00, sha 017cd7da2239)
  - homepage: https://jianhongbai.github.io/ReCamMaster/ (fetched 2026-08-29T10:55:47.334986+00:00, sha 22ef825ee85c)
- Data as of 2026-08-30T08:39:29.467469+00:00.