Deploy guides
Docker image to public URL
Each guide lists the image, container port, health endpoint and the smallest GPU that fits — plus the cheapest card that meets that floor right now, from the live Vast.ai market.
vLLM
Serve an OpenAI-compatible chat completions API from a Vast.ai GPU, with one public URL, health checks and automatic replacement when an instance dies.
- Port
- 8000
- Min VRAM
- 24 GB
- Fits from
- $0.10/hr
Ollama
Run an Ollama server on a rented GPU and reach it over HTTPS, with automatic model pulls on boot, health checks and instance replacement.
- Port
- 11434
- Min VRAM
- 16 GB
- Fits from
- $0.09/hr
ComfyUI
Serve a ComfyUI instance on a Vast.ai GPU behind one URL, so image and video workflows can be called from your app instead of from a desktop.
- Port
- 8188
- Min VRAM
- 16 GB
- Fits from
- $0.09/hr
Any Docker image works
These three are the common ones. CUDAnetes deploys any container that serves on a port and answers a health path: set the image, the port and the endpoint, pick your GPUs, get one public HTTPS URL.