The people-powered inference network
AI, powered by all of us.
weandi spreads AI work across thousands of everyday devices: laptops, desktops and GPU rigs running open models in a locked-down sandbox. Builders get affordable inference. The people who power it get paid.
How it works
-
1
You send a job
One prompt, a batch of thousands, or a long document. Same simple API you already know, with a clear price up front.
-
2
We split it up
weandi breaks the work into pieces and matches each one to a device that can handle it, from laptops to multi-GPU rigs.
-
3
Everyone gets paid
Results come back verified. The people whose devices did the work earn for every piece they complete.
For builders
Affordable inference, without a GPU contract
- Open models, priced per model, in plain dollars
- Built for batch jobs, long documents and many samples
- Familiar, OpenAI-style API
- Random spot-checks keep results honest
For providers
Your idle computer can earn
- Install the app, choose when your device works
- Models run fully sandboxed: no internet, no file access, no tools
- Set the minimum you're willing to accept
- Earnings shown in dollars, cashed out to your bank
What we believe
Power to the people
AI capacity shouldn't belong only to a few data centers. Millions of capable devices sit idle every night.
Fair by design
Same price for the same work. Faster devices earn more by doing more, not by gaming the system.
Safe to run
Every model runs in a sealed sandbox. Your device does inference, nothing else.
Simple money
Dollars in, dollars out. Payments are recorded on a public ledger behind the scenes, so you never have to think about it.
Request early access
We're onboarding builders and device providers in small groups. Tell us a little about yourself and we'll be in touch.
Questions
Is my data safe when it runs on someone else's device?
Models run in a sandbox with no network or file access, and results are spot-checked. Prompts are processed on provider devices, so for sensitive workloads we're building a private tier that runs only on hardware-isolated or vetted machines.
Do I need crypto or a wallet?
No. You pay and get paid in dollars. A public ledger records payments behind the scenes for transparency; you never have to touch it.
What devices can provide compute?
Modern laptops and desktops (Apple Silicon Macs, Windows and Linux machines with GPUs) and dedicated GPU rigs. Smaller devices run smaller models; bigger machines take on bigger ones.
Which models are supported?
Popular open-weight models in GGUF format, from small fast models to large ones. The launch lineup will be announced to early-access members first.