Documentation
Getting Started
- Create an account and subscribe ($10/month)
- Click "Create Desktop" on your dashboard
- Save your password (shown only once!)
- Click "Start Desktop" when you want to connect
- Use any RDP client to connect
Connecting via RDP
Windows
Use the built-in Remote Desktop Connection (mstsc.exe)
macOS
Download Microsoft Remote Desktop from the App Store
Linux
Use Remmina, FreeRDP, or any RDP client
Connecting via SSH
Use any SSH client with the credentials shown on your dashboard:
ssh user@host -p PORT
Public Ports
Each user gets 10 HTTPS-accessible ports. Run a server on any port inside your container and access it via:
https://fits.net/public/YOUR_PORT/
Perfect for exposing APIs, web apps, Jupyter notebooks, or any other service.
AI Workstation Plan
$10/month
- ✓ 2 CPU cores
- ✓ 4 GB RAM
- ✓ 4 GB GPU VRAM
- ✓ 25 GB free storage
- ✓ Unmetered access
- ✓ 10 public HTTPS ports
Subject to fair use policy. Cancel anytime.
Fair Use Policy
Unmetered access is for interactive AI/ML work:
- Running LLM inference (Ollama, llama.cpp)
- Embeddings and RAG applications
- Development, Jupyter notebooks, research
- Speech recognition (Whisper), embeddings
Not permitted:
- Sustained 24/7 max CPU/GPU load
- Cryptocurrency mining
- Unattended batch processing for days
- Sending email/spam (ports blocked)
Note: This is a compute-only GPU sandbox. Graphics (OpenGL, Vulkan) and video encoding (NVENC) are not available.
See full Terms of Service for details.
Important: Spot Instances
This service uses spot/interruptible GPUs
- Server may be interrupted at any time without warning
- Your files are safe on persistent storage
- Just reconnect when the server is back
- Save your work frequently!
Tips
- Save your work frequently - spot instances can be interrupted
- Your files persist even when stopped
- Stop your desktop when not in use to free resources for others
- Use public ports to expose web services and APIs
Pre-installed Software
AI / ML
- PyTorch 2.9 + CUDA 12.8
- Python 3.12
- JupyterLab
- NumPy, Pandas, scikit-learn
Tools
- nvidia-smi
- Build essentials (gcc, make)
- Git
- Docker (optional)
Desktop
- Ubuntu 24.04 LTS
- Chrome browser
- Full sudo access
- RDP + SSH access