Repository files navigation

Cubewave serves as the Heroku for AI agents, allowing easy deployment of your LLM in a secure cloud environment from tools like Rivet, langflow...

Cubewave is providing tools to make deploying microVMs through Firecracker simpler, minimizing the complexities entailed in the process.

• Corewave operates as the equivalent of "docker-compose" for microVMs, enabling the configuration and creation of microVMs through a YAML file.

To keep updated on our progress, join our discordserver and follow us onTwitter!

In order to create micro vm, you need to make sure KVM is enabled on your machine

lsmod | grep kvm

The output should be something like this:

kvm_intel             342150  0kvm                   960692  1 kvm_intelirqbypass              15324  1 kvm

Cubewave currently runs on Linux Ubuntu and you also need to have at least 5GB of available storage

    wget --quiet -O - https://repo.cubewave.dev/cubewave-pubkey.asc | sudo tee /etc/apt/keyrings/cubewave-pubkey.asc     echo "deb [signed-by=/etc/apt/keyrings/cubewave-pubkey.asc arch=$( dpkg --print-architecture )] https://repo.cubewave.dev $(lsb_release -cs) main" | sudo tee /etc/apt/sources.list.d/apt-cubewave.list    apt update    apt install corewave

To verify that corewave is successfully installed runcorewave version

you should get something like this:corewave v.0.1

Corewave requires a YAML files that defines the configuration of the MicroVM(s) you want to deploy, below is how the file should be structured:

version:"1"specs:MyApp1:logsPath:"FC_MyApp.log"CPU:1RAM:1024image:"sample_app"MyApp2:...MyApp3:...

version: Configuration version

spec: Specs block, within we specify the configuration of all our MicroVMs

MyApp1: Name of the MicroVM to provision (you can name it whatever you want)

logsPath: Path of the Firecracker log file

CPU: Number of host CPUs to allocate to the MicroVM

RAM: Amount of RAM to allocate to the MicroVM (in MB)

Image(optional): Absolute path to any executable you want to run as a daemon in the MicroVm.

Once your YAML file is ready you run Corewave and pass it the configuration file using the-f flag:

sudo corewave -f config.yml

Corewave will then generate the necessary infrastructure components on your host VM and run firecracker with the required configuration to provision your MicroVMs.

Note: For now Corewave needs to run with privileged permissions so please make sure you are either root or using sudo when running Corewave

/!\After each new executing of Corewave to provision new MicroVMs, the previous set of MicroVMs and data will be deleted. This will be changed in the upcoming release. /!\

Once your MicroVMs are up and running, Corewave provides a set of commands to easily monitor and interact with your MicroVMs:

Corewave stores all data and configuration related to each MicroVM in the default$HOME directory under.cubewave

Usecorewave ps to get the list of currently running MicroVMs

You can also use the ps argument to get information about a specific MicroVM only, by providing its id using corewave ps {id}

You can simply usessh to connect to a MicroVM, authentication is based on an ssh key that located in the config directory of every MicroVM under$HOME/.cubewave/{ID}/fs/ubuntu-22.04.id_rsa

{ID} is ID of the MicroVM you want to connect to.