Proof of Use. An independent field guide · No. 01
Back to the collection
3D

Nerfstudio

A toolkit for turning overlapping photos or video into trainable 3D scene representations and rendered views.

3D reconstructionNeRFGaussian splattingphotogrammetryPython
Source-reviewedNot hands-on testedSources checked

Start with an idea

What you can make

01

Capture an object from several angles, train a scene and render a smooth orbit video.

02

Reconstruct a captured space for a new camera-path walkthrough or an exported Gaussian splat.

Creator demonstration

See the documented workflow

The official page embeds a tutorial using the legacy viewer. It explains that the current interface differs while the core workflow remains similar.

Creator capture-to-video tutorial

Before you begin

What it takes to run

Python/CUDA + capture processing

  • Requires a Python/PyTorch/CUDA environment and GPU-related dependencies. Photo/video preprocessing also uses COLMAP and FFmpeg.
  • The Splatfacto docs estimate about 6 GB memory for the default model and 12 GB for the larger variant; scene and settings affect actual needs.
Runs through / on

Linux · Windows with documented build setup

The real tradeoffs

Cost & licensing

Running costs

No core software license fee; GPU training and capture costs apply.

The documented default route needs NVIDIA/CUDA hardware. Rented compute, storage and optional capture apps can add costs.

Code license

Apache-2.0

Read the code license

Models & other terms

No single bundled model license covers every method or dataset

Check any optional method, dependency, capture app and dataset separately; the root license is for Nerfstudio code.

Know the limitations

  • Good overlapping, non-blurry captures matter; camera-pose estimation can be finicky.
  • A radiance field or Gaussian splat is not automatically a clean editable mesh; the documented Splatfacto exporter does not provide mesh export.

Follow the evidence

Sources & verification

Official sources were reviewed on October 14, 2026. We did not install, run or benchmark this project. This is a source-check date, not a release date.

Links and project terms can change. These sources document the claims above; they are not a safety certification. Read our full method.

Keep exploring

A few other possibilities

All tools

09Images & Video

ComfyUI

A visual node editor for building repeatable AI image, video and other media workflows.

Try it for

Build a reusable image-generation workflow, then repaint selected areas with an inpainting model.

Read entry Official examples
Setup · Moderate
Desktop app + model workflow
Cost to consider
Local compute; optional paid cloud/API

10Images & Video

Diffusers

A Python library for wiring diffusion models into image, video and audio applications.

Try it for

Prototype a text-to-image feature in your own app with a selected model and repeatable parameters.

Read entry Official examples
Setup · Advanced
Python + model pipeline
Cost to consider
Local or rented compute

11Images & Video

SAM 2

A promptable segmentation model that produces object masks in images and tracks them through video.

Try it for

Create a moving-subject mask for a video-editing or annotation pipeline, then review and refine it.

Read entry Official examples
Setup · Advanced
Python + CUDA/checkpoints
Cost to consider
Local or rented GPU compute