Meet Cannabis. Put ideas to work.

Meet Cannabis, a community built around a simple idea: putting computing power to useful work.

Meet Cannabis, a community built around a simple idea: putting computing power to useful work.

The project brings together open code, clear rules and privacy-conscious design, with a community allocation built around people.

As we prepare the public launch, you can explore the source, contribute your skills and help shape what comes next. We look forward to building it together.

You can help Cannabis grow in several ways, from reviewing code and sharing original artwork to asking questions that improve the project.

The official website and GitHub repository are good places to start. There, you can read the published rules, follow release updates and join the discussion.

When enrollment opens, follow the verified instructions and use your own wallet, keeping your recovery secrets private. For now, you can explore the project and find a way to contribute.

Cannabis is designed for efficient participation, without a proof-of-work mining race or a competition to spend computation on new tokens.

The target network uses Ethereum proof-of-stake settlement. Useful computation happens separately, through tasks that requesters need and agree to pay for.

These tasks can include training, rendering, analysis, simulation and other software workloads. The aim is to put computing resources to productive use, with measured costs and clear results.

The Cannabis compute client uses a general-purpose workload contract. Package your software, declare its runtime, specify inputs and outputs, and set resource limits.

New adapters can extend the format. A script, a virtual machine checkpoint, a renderer, or coordinated model training can each describe what they need.

Whole jobs can target one suitable node. Workloads with declared dependencies can form a task graph. Distributed programs can request compatible groups. The runtime and splitting rules preserve what your program is meant to do.

Your code, inputs, checkpoints and outputs belong to your job. Cannabis requires encryption for stored files and network transfers, as well as protection while your program runs.

The execution design releases keys only into an attested environment that meets the job's policy. Each participating accelerator and network path must support that protection.

Ordinary hardware compatibility alone is not enough. The confidential worker service still needs integration and review. Public payments and some scheduling metadata remain visible. Explore the design and help strengthen it.