3D Gaussian Splatting
The INRIA original — train your own splats.
Open Source 12–24 GB VRAM
- Min VRAM
- 12 GB
- GPU class
- Mid GPU
- Quant
- FP32
Actually FreeNo SignupWatermark-Free
Nerfstudio
The open framework for NeRF and Gaussian Splatting research.
Open Source 8–24 GB VRAMRuns locally
Actually FreeNo SignupOpen SourceWatermark-Free
Visit NerfstudioUpdated 2026-04-30 · Direct link
Hardware requirements
Runs locally · Mid GPU (12 GB)
Min VRAM
8 GB
Rec. VRAM
24 GB
Min RAM
16 GB
Rec. RAM
32 GB
Disk
30 GB
GPU class
Mid GPU
11.8+No Apple SiliconGPU RequiredQuant: FP32
Splatfacto on 12 GB; full pipeline comfortable at 24 GB.
Screenshot placeholder · Nerfstudio
What is Nerfstudio?
Nerfstudio is the standard OSS pipeline for training NeRFs and 3D Gaussian Splats from photo/video input. Modular trainers (Nerfacto, Splatfacto, Instant-NGP) all under one CLI. Used heavily in research — your reference if you want to train your own representation rather than use a SaaS.
Pros & cons
Pros
- Reference implementation for NeRF + Gaussian Splatting research
- Modular: swap trainers, samplers, datasets via config
- Excellent web viewer for inspecting trained scenes
- Active academic community
Cons
- Setup is non-trivial; CUDA / colmap dependencies have sharp edges
- Not aimed at end-users — expect to read papers, not tutorials
What's actually free?
Apache 2.0 from UC Berkeley KAIR-BAIR.
✓ Actually FreeNo SignupOpen SourceWatermark-Free
Alternatives
Instant-NGP
NVIDIA's seconds-to-train hash-grid NeRF.
Open Source 8–12 GB VRAM
- Min VRAM
- 8 GB
- GPU class
- Entry GPU
- Quant
- FP16
Actually FreeNo SignupWatermark-FreeHobbyist-Friendly
Luma AI
Phone-scan to NeRF, Genie text-to-3D, and Dream Machine video.
Freemium · from $30/mo☁ Cloud · no GPU
Hobbyist-FriendlyAPI