RunSnack is an innovative tool designed to make GPU sharing as simple as sharing a single link. Developed by AI researcher Manuel Herrador Muñoz, PhD, this platform allows users to share their hardware resources with anyone, anywhere, effortlessly.

For hosts, sharing a GPU is as simple as running a script to generate a unique link. For renters, opening the link provides an instant GPU terminal directly within their browser. Security and privacy are prioritized through sandboxed Docker containers that ensure zero host access, alongside zero-knowledge P2P WebRTC communication that ensures no central server accesses your data.

Operating with zero platform fees, RunSnack allows hosts and renters to negotiate terms directly. The marketplace and community—where users can post offers, find available GPUs, and build their reputation—are hosted directly on Discord.

Pros of RunSnack — Your GPU. One link. The world

Share or access a GPU instantly via a single link
Zero platform fees with direct peer-to-peer negotiation
Enhanced security through sandboxed Docker containers with no host access
Zero-knowledge architecture using P2P WebRTC to protect user data

Cons of RunSnack — Your GPU. One link. The world

The marketplace relies entirely on Discord for finding GPUs and building reputation
Requires running a setup script for hosts to generate links

Frequently Asked Questions

How do I share my GPU using RunSnack?

As a host, you simply run a script to generate a unique link which you can then share with anyone.

How do renters access the GPU?

Renters simply open the provided link to get an instant GPU terminal directly in their web browser.

Are there any platform fees?

No, RunSnack features zero platform fees, allowing users to negotiate terms directly with one another.

How is security and privacy maintained?

RunSnack utilizes sandboxed Docker containers so renters have no host access, alongside a zero-knowledge P2P WebRTC architecture ensuring no central server sees your data.