33 ready software setups. Your token’s GPU boots with the template already installed, then its AI builds on it the way you describe.
Template for NVIDIA RAPIDS data science toolkit.
A clean GPU toolkit for building anything custom from scratch.
PyTorch for Nvidia CUDA
Nvidia CUDA + CuDNN development base image.
Unsloth Studio LLM training and inference
Train or fine-tune your own AI model on custom data.
Llama.cpp
Run your own AI model: a chatbot, assistant or API for apps and bots.
ComfyUI image and video generation software
Generate images and art from text: galleries, avatars, collections.
vLLM inference engine
vLLM Omni inference engine
ComfyUI, SD Forge, ACE Step, Voicebox, Whisper, Wan2GP, AI Toolkit and Unsloth Studio.
Create AI videos and clips from text or images.
ACE Step 1.5 music generation interface and API
Speech-to-text, AI voices or AI music.
Official images for the TensorFlow deep learning framework.
PyTorch NGC container with DALI, RAPIDS, cuDNN, NCCL and TensorRT.
AI toolkit for training diffusion models
Kohya_ss GUI for Stable Diffusion and FLUX training
Flux Gym for training FLUX.1 image model LoRAs
Voicebox TTS
Whisper web interface and API service
Axolotl streamlines architecture, datasets, configuration and GPU access.
Ubuntu Linux desktop container plus the Pinokio installer.
A full Linux desktop with one-click AI app installers.
Text generation UI and API
Ubuntu Linux desktop container
SGLang for inference
Ollama template with configurable model loading
Open WebUI with an Ollama backend
Text Generation Inference server
Langflow with Ollama
InvokeAI image generation
Forge image generation for Stable Diffusion and FLUX
Wan2GP video generation
Open-Sora text and image to video model
A1111 image generation for Stable Diffusion
Fooocus image generation
SwarmUI image generation interface
Select a template above, or launch without one.
fior