Your graphics card can pay for itself.
Install Engine, pair your RTX, and Grid sends it AI jobs — chat, images, voice, video — whenever you're not using it. Every finished job adds to your balance. Bank-transfer payouts are opening soon.
What does my GPU earn?
70% of eligible job revenue back to youMonthly earnings simulation
- Earnings (completed jobs only)
- $13.30
- Electricity (450 W × hours, upper bound)
- −$9.18
- Estimated net
- $4.12
Job allocation is not guaranteed. How many jobs you receive depends on network demand and your GPU’s availability, and some months may bring none. This is an estimate from the assumptions you entered, not a promise of income.
How it is shared: you receive 70% of the public API price (in credits) billed to the requester for each completed job. Jobs that fail — errors, timeouts — are not billed to the requester and earn nothing. Earnings from jobs flagged by quality checks are held until the review is done. (1 credit = $0.001)
Any RTX 40 or 50 series card. From 8 GB up.
Grid reads your VRAM and only sends jobs your card can finish. More memory unlocks bigger models — and better-paying work.
Older cards (RTX 30, 20) and macOS: coming later. Laptop GPUs work if they meet the VRAM tier.
Three steps. About five minutes.
Install Engine
Download OpenAlchemy Engine for Windows. It finds your RTX card and checks your driver.
Pair your GPU
Sign in, copy a one-time code from your dashboard and paste it into Engine.
Get paid
Every finished job adds to your balance. Add your bank account now — payouts are opening soon.
Game first. Earn the rest of the time.
You decide when it works and when it stops. Auto-pause for games and power caps are on the way.
Auto-pause for games
coming soonLaunch a game or a GPU-heavy app and Engine steps aside.
Set your hours
Share only overnight, only on weekends, or all day.
Power cap
coming soonLimit wattage and temperature so your card runs cool and quiet.
Live dashboard
See every job, your uptime, your balance and your payouts, as it happens.
Open-source models, served by the network.
Grid picks the models your card can hold and routes matching jobs to it. Nothing to configure.
| MODEL | TYPE | MIN VRAM | STATUS |
|---|---|---|---|
| wan2.2-i2v | video · image to video | 24 GB | live |
| wan2.2-s2v | video · singing to video | 24 GB | live |
| wan2.2-animate | video · motion transfer | 24 GB | live |
| latentsync-1.6 | video · lip-sync | 24 GB | live |
| qwen-image-2512 | image | 16 GB | live |
| qwen-image-edit-2511 | image · edit | 16 GB | live |
| qwen3.5-9b | chat | 12 GB | live |
| cosyvoice3 | text-to-speech · voice cloning | 8 GB | live |
| sensevoice | speech-to-text | 8 GB | live |
| bge-m3 | embedding | 8 GB | live |
Questions
How do I get paid?+
Earnings build up in your provider dashboard as your GPU finishes jobs. Bank-transfer payouts are opening soon: add your bank account in the dashboard now, and once payouts open you can cash out from 5,000 credits ($5.00), paid within 10 business days. Until then you can spend earnings on the OpenAlchemy API.
Will it slow down my games?+
Automatic pause for games is coming soon. Until then, stop sharing in Engine before you play and start it again when you are done.
What about my electricity bill?+
Power and temperature caps are coming soon in Engine. Earnings above are before electricity; check your local rate to see your net.
Is my computer safe?+
Engine only runs signed OpenAlchemy model workers. Jobs never get access to your files, and you can uninstall at any time.
Which GPUs are supported?+
Every GeForce RTX 40 and 50 series card, desktop or laptop, with 8 GB of VRAM or more. Support for RTX 30 series is planned.
Your GPU is idle right now.
Put it to work.
Windows 10/11 · NVIDIA driver 530+ · RTX 40 or 50 series
Running Linux? Get the command-line worker →