Get started

Connect your wallet

Open the console and connect Phantom. Your Solana wallet is your identity, your balance, and how you sign deployments — no email or signup.

Add $PLOS to your balance

Compute is paid in $PLOS, metered per second. Top up from Billing → Add funds, or enable auto-reload to refill when your balance runs low. You only ever pay for completed work.

Indicative cost: an RTX 4090 is ~0.34 PLOS/GPU-hr, so a 20-minute, 2-GPU render runs ≈ 0.23 PLOS.

Deploy a job

Pick a Marketplace template (or bring your own Docker image), choose a GPU and region, set a budget, and deploy. The Core System decomposes the job into work units and routes them to matching nodes automatically.

# in the console
Marketplace → choose "ComfyUI · SDXL" → Deploy
   ↳ GPU: RTX 4090 ×2 · region: us-east · budget: 40 PLOS
   ↳ input: ipfs://bafy…/refs.zip

parallelos deploy \
  --template comfyui-sdxl \
  --gpu "RTX 4090" --count 2 --region us-east \
  --budget 40 --input ipfs://bafy…/refs.zip

job = client.deployments.create(
    image="docker.io/parallelos/comfyui:2.4",
    gpu="RTX 4090", count=2, region="us-east",
    inputs="ipfs://bafy…/refs.zip", max_budget=40,
)

Watch it run

Follow the live log stream as the image is pulled, inputs are mounted and work units execute across nodes. Progress and per-unit status update in real time.

Collect your results

On completion the outputs are assembled and uploaded as downloadable artifacts. Grab them from the deployment's Results tab, or pull them with the CLI. Fees settle automatically — failed units are refunded.

parallelos pull dep_7a3f9c -o ./out
# → renders_batch.zip (248 MB), grid_preview.png, run.log

That's it — you've run a job on decentralized GPUs. Explore more templates in the Marketplace or automate with the API & SDK.

Check your hardware

Any modern NVIDIA GPU with 8 GB+ VRAM can contribute. More VRAM, compute and stable bandwidth mean larger units routed your way — and higher earnings.

Install the Agent

Run one command on the machine you want to contribute. The Agent connects outbound only — you never expose any ports.

curl -fsSL https://get.parallelos.network | sh

# PowerShell (admin)
irm https://get.parallelos.network/win | iex

docker pull parallelos/agent:latest

Pair your device

Open Connect Device in the console to get a one-time pairing token, then start the Agent with it. The token links this machine to your wallet.

parallelos agent start --token plos_xxxxxxxx_xxxxxxxx

Verify it's online

The Agent runs a one-time benchmark. Your device appears in Devices as pending, then online once verified, and hired while running a job.

Earn & claim $PLOS

Rewards accrue per device from completed work, performance and uptime. Watch live telemetry on the device page, and claim your balance to your wallet anytime from Earnings. Consistent uptime raises your reputation and scheduling priority.