IDOL (Image-to-3D DiT)
Speridlabs' open-source 3D Diffusion Transformer (3D-DiT). Converts single photos into dense 3D Gaussian Splats (up to 1,048,576 Gaussians) outputting standard .ply files under an Apache 2.0 license.
The brutal truth
The most practical open-weight 3D Gaussian generator today. Requiring only ~7.2 GB VRAM under Apache 2.0 makes high-fidelity 3D spatial generation accessible on standard laptops and desktops.
Files and Versions
Direct inspected tree artifacts from Hugging Face. Verified filenames, exact byte sizes, model weight shards, and configuration definitions.
Decision matrix
✓ RUN / USE THIS IF…
- •You own an 8GB or 12GB consumer GPU and want zero API fees
- •You want an Apache 2.0 open-weight model with ComfyUI node workflows
- •Your 24GB GPU comfortably fits the model weights (~6GB VRAM) with dedicated headroom for inference.
- •You require 100% offline data privacy with zero third-party API logging, subscriptions, or prompt retention.
✕ WALK AWAY IF…
- •You require quad-mesh polygon topology with UV maps (use Tripo/Meshy instead)
- •Your system has less than 6GB dedicated VRAM
- •Your GPU has less than 6GB VRAM and you cannot tolerate slow CPU/system RAM offload paging.
- •You prefer a plug-and-play web app interface and don't want to manage local model checkpoints, drivers, or runtimes.
True cost of ownership
About
Speridlabs' open-source 3D Diffusion Transformer (3D-DiT). Converts single photos into dense 3D Gaussian Splats (up to 1,048,576 Gaussians) outputting standard .ply files under an Apache 2.0 license.
What it does
Download the weights from Hugging Face, pick a quantisation that matches your GPU, and run it locally with Ollama, llama.cpp, vLLM, or a compatible runtime. Single-image to 3D Gaussian Splatting runs on your own hardware — no per-token bill, no third-party data handling.
Key features
Converts single 2D images into dense 3D Gaussian Splats (.ply) in seconds
Runs on consumer graphics cards (~7.2 GB VRAM requirement)
Permissive Apache 2.0 license with interactive Spaces and ComfyUI integration
Use cases
Turn single photographs into dense 3D Gaussian Splats in seconds.
Create interactive 3D product previews for web stores without studio capture.
Generate game-ready spatial assets and environment props directly from concept art.
Export standard .ply assets ready for WebXR, Vision Pro, Unity, and Unreal Engine.
Quantization & Hardware Fit
VRAM Hardware Compatibility & Memory Allocation
| Quant | VRAM | Quality / Fit on 24GB | Install / Run Command (OLLAMA) | |
|---|---|---|---|---|
| FP8● ACTIVE | 6 GB | Standard · ✓ Fits (18GB spare) | ollama run speridlabs-idol | |
| BF16 | 8 GB | Standard · ✓ Fits (16GB spare) | ollama run speridlabs-idol | |
| FP16 | 12 GB | Standard · ✓ Fits (12GB spare) | ollama run speridlabs-idol |
Buyer doubts, cleared
Can I actually run IDOL (Image-to-3D DiT) on my 24GB GPU?
What's the fine print on the license?
Does the 1024×1024 Single Image actually hold up in practice?
User reviews
Reasoning quality is close to the bigger models for structured tasks. Open-ended planning still needs a frontier model.
Solid for single-image to 3d gaussian splatting. The learning curve is real but the payoff is there. Start with the smaller build first.
We evaluated IDOL (Image-to-3D DiT) against two hosted alternatives. It won on cost-per-output but lost on latency. For batch work it is a clear yes.
Switched our internal workflow to IDOL (Image-to-3D DiT) after the last pricing change. Nobody noticed a quality difference, and the bill dropped by half.
