Getting Started
Install 360 Hextile, activate your license, and create your first 360° panoramic image
Last updated: 2026-10-27
Windows 10 (64-bit, 21H2+) or Windows 11 and a CUDA-capable NVIDIA GPU. Check your card
Installation
- Purchase a license at 360hextile.com/pricing
- Download the installer from your account dashboard
- Run the installer and follow the on-screen prompts
- Launch 360 Hextile from your Start menu or desktop shortcut
First-Run Wizard
On first launch, 360 Hextile walks you through a setup wizard:
- Welcome — Product introduction
- Storage — Choose your data root folder (models, configs, outputs)
- License Agreement — Read and accept the license agreement
- License — Activate online using your account or license key
- GPU — Automatic GPU detection and VRAM check
- Telemetry — Optional, off unless ticked; licensed totals go to your login-only 360hextile.com dashboard — not public, not Gallery, never prompts or images
- Models — Select AI models to download (SDXL recommended to start)
- Complete — Ready to use
After setup, use Settings for the corresponding preferences. See License Activation for the separate activation task.
HuggingFace Token for Gated Models
Some models (like SD 3.5) are gated and require HuggingFace authentication:
- Create a free account at huggingface.co
- Accept the model license on the model's HuggingFace page (e.g., Stability AI's license for SD 3.5)
- Generate an access token at huggingface.co/settings/tokens
- In 360 Hextile, go to Settings and paste your token in the HuggingFace section
- Save the token and check the selected model's access result. A valid token does not grant access to every gated repository.
See Model Management for more details.
GPU Detection
The application automatically detects your NVIDIA GPU and VRAM. The status bar shows your available GPU memory.
Your First Render
Step 1: Start Your Panorama
For a straightforward first task, open Create → Import Image, choose a JPG/PNG equirectangular panorama (2:1), inspect its preview and click Import. See Import Image. If you have no image, 360 LoRA generates panorama variations after its own model setup; Pattern or Gradient needs no diffusion model to create a starting image.
Step 2: Choose a Template
Start with Dodecahedron and a modest image/workload. More faces or larger dimensions are later choices, not prerequisites for first success. See Templates.
Step 3: Configure Your Settings
Choose an installed model suited to your hardware. Write a short environment prompt and keep the selected model's default steps/guidance for the first comparison. Denoising strength changes how far the input is transformed; lower it for a more conservative test. Numeric schedules and negative-prompt support differ by model: see Parameters.
Step 4: Process
Click RENDER and inspect that job's progress. Wait for completion or a clear failure/cancellation state before starting replacement work. Generation time depends on model, hardware and settings. See Troubleshooting if it fails.
Step 5: Export
Inspect the completed result in the sphere, including seams and poles, then use Exporting Images for a PNG/JPG copy at a practical size. HDR/float output, projection conversion and Shader Linear EXR have distinct requirements. Check the exported file in your receiving workflow before scaling up.
Tips for Best Results
- Start with Dodecahedron — it is the default factory template
- Compare strength on a small test — the useful range depends on the model and the change you want
- Use directional prompts for directional variation (different sky, ground, etc.)
- Try different AI models — each has its own artistic character
- Use compatible ControlNet when the selected pipeline supports the structural guidance you need
Next Steps
- Interface Overview — Learn the full application layout
- Pipelines — Understand different AI models
- Templates — Choose the right template
- Modules — See which optional features each pipeline supports
- ControlNet — Depth-guided generation
- Settings — Configure preferences
- Learning Mode — Optional bundled guidance
- Rendering and Recovery — Follow queued work and inspect uncertain outcomes
Need help? Check the Troubleshooting Guide or contact us.