For GPU Contributors
Put your GPU to work.
Most gaming and workstation GPUs run at under 10% utilization for the majority of the day. The Sogni provider client uses that idle headroom to run image and video inference jobs — you earn per completed job, the scheduler only assigns work when your GPU is not otherwise busy.
How to contribute
Three steps to start earning
1. Install provider client
Download and install the Sogni provider client for Linux or Windows. Lightweight background service — minimal CPU overhead, uses your GPU only during jobs.
2. Configure your GPU specs
Enter your GPU VRAM, set your availability hours and maximum concurrent jobs. The scheduler only assigns jobs compatible with your hardware.
3. Earn per completed job
Credits accumulate for every generation job your node completes. Monitor earnings in your dashboard. Redeem monthly via your preferred payment method.
Requirements
Hardware requirements
| Requirement | Minimum | Recommended |
|---|---|---|
| GPU VRAM | 8 GB | 16 GB or more |
| Operating System | Windows 10, Linux (Ubuntu 20.04+) | Linux preferred for server deployments |
| Internet connection | Stable 50 Mbps+ | 100 Mbps+ with low latency |
| Storage | 50 GB free (model cache) | 200 GB+ for multiple model variants |
| Uptime | Any availability | 8+ hours/day for higher job volume |
| GPU driver | CUDA 11.8+ or ROCm 5.4+ | Latest CUDA / ROCm stable |
Earnings
Your compute, your income
Earn credits for every inference job your node completes. Credit weight reflects the model's VRAM requirement and the actual compute time consumed — a FLUX 1024×1024 generation earns more than a SD 1.5 512×512 one. Credits accumulate in your dashboard and are redeemable monthly.
- No lock-in period — connect or disconnect anytime
- Uptime bonus credits for high-availability nodes
- Real-time earnings dashboard
- Monthly redemption via supported payment methods
- No token or cryptocurrency required — straightforward credits
Provider FAQ
When do I get paid?
Credits are redeemable at any time from your dashboard. Monthly redemption cycles ensure stable payouts.
Will jobs affect my gaming performance?
The provider client monitors GPU utilization and holds jobs when you're gaming or rendering locally. You can also set a hard pause from the dashboard or system tray icon. Jobs are not queued for your node when paused — they route elsewhere.
What kind of jobs run on my GPU?
Only diffusion and video inference — no training runs, no data storage on your disk, no cryptocurrency mining. Model weights are cryptographically verified before execution and are not retained after the job completes.