1 GPU online of 3 listed, by people across Saudi Arabia

GPUs in Saudi Arabia.By the hour.

A marketplace, not a data center: people list their idle NVIDIA cards in Riyadh, Jeddah, or Dammam, and you rent them by the minute in SAR.

Join the waitlist I have an invite

Invite-only beta. Starter credits, no card, no real charges.Have an idle GPU? List it and earn

llama3.2:3b answered on Cooper's 3090 · 11.4 s14 h ago
llama3.2:3b answered on Cooper's 3090 · 12.4 s14 h ago
llama3.2:3b answered on Cooper's 3090 · 13.6 s15 h ago
llama3.2:3b answered on Cooper's 3090 · 33.8 s15 h ago
Cooper's 3090 came online19 h ago

Available nowSee all 3 nodes

GPUVRAMCityModels readySAR/hr

NVIDIA GeForce RTX 3090Online24 GBRiyadhllama3.2:3b, qwen2.5:7b, nomic-embed-text10Rent

NVIDIA GeForce RTX 3090Offline24 GBRiyadhllama3.2:3b, qwen2.5:7b, nomic-embed-text12Rent

NVIDIA Jetson AGX Orin Developer KitOffline61 GBRiyadhqwen2.5:7b, llama3.2:3b, phi312Rent

NVIDIA GeForce RTX 3090Online

24 GB · Riyadhllama3.2:3b, qwen2.5:7b, nomic-embed-text

10 SAR/hrRent

NVIDIA GeForce RTX 3090Offline

24 GB · Riyadhllama3.2:3b, qwen2.5:7b, nomic-embed-text

12 SAR/hrRent

NVIDIA Jetson AGX Orin Developer KitOffline

61 GB · Riyadhqwen2.5:7b, llama3.2:3b, phi3

12 SAR/hrRent

01Real cards, owned by real peopleEvery GPU is a physical NVIDIA card in someone's machine inside Saudi Arabia: a studio, a lab, a gamer's rig. Your data never leaves the country.

02Priced by the minute, in SARTake a card for ten minutes or hold a whole node for days. The total is shown before you confirm. End early and pay only for what you used.

03Any workload, two ways inCall ready models through an OpenAI-compatible API, or take the whole GPU for training and fine-tuning with your own stack.

FOR BUILDERS

Use open models without buying a GPU.

Whether you want to try the latest open-source models or ship an agent, you pay for what you use: by the token for API calls, by the minute for a whole card. No hardware, no commitment, nothing idle on your bill.

Explore open models

Llama, Qwen, Mistral, DeepSeek and Arabic-tuned models like Aya. Pick one, chat, compare. Starter credits cover your first experiments.

Chat in the browser, no setup
Switch models mid-conversation
Arabic-tuned models served from inside the Kingdom

Start with free credits

Build your agent

An OpenAI-compatible endpoint your code already understands. Tools, JSON output and streaming work with LangChain, the Vercel AI SDK, CrewAI and n8n.

Same base_url and api_key, nothing to rewrite
Pay per token; hold a whole GPU only when you need it
Your prompts stay on cards inside Saudi Arabia

Start with free credits

PROVIDERS

Your idle GPU should be earning.

One command on a Linux machine with an NVIDIA card. The agent dials out to Sadeem, never the other way around. Renters run models through the platform and never get a shell on your machine.

Outbound connections only, no open ports
You set the price per hour and the model list
Earnings in SAR, tracked by the minute

List my GPU

# linux + nvidia gpu
$ curl -fsSL https://api.sadeem.run/provider/install.sh | sh

gpu      RTX 4090, 24 GB
models   3 ready
status   riyadh-4090-01 online and listed

EARLY ACCESS

We are opening Sadeem to a small group first.

Leave your email and what you want to do. We invite people in batches, work with them on the product, and open up when it is ready.

Renters get starter credits; no card, no real charges during the beta.
Providers get set up with us on a call and see earnings from day one.
A short confirmation now, one email when your invite is ready. Nothing else.