Minikube (mini-langstream)

To create a LangStream cluster locally, it's recommended to use minikube - setting 4 CPUs is highly recommended. mini-langstream comes in help for installing and managing your local cluster.

Install mini-langstream

mini-langstream requires the following commands to be already installed:

• Docker

• Minikube

• Helm

• Kubectl

• LangStream CLI

If you install mini-langstream using Brew, all the dependencies are automatically installed.

MacOS:

brew install LangStream/langstream/mini-langstream

Unix:

curl -Ls "https://raw.githubusercontent.com/LangStream/langstream/main/mini-langstream/get-mini-langstream.sh" | bash

Start the cluster

mini-langstream will do all the setup for you, in particular:

• start minikube in a dedicated context

• deploy LangStream components using helm

• run a stateful Kafka broker as docker container

• run a stateful s3-compatible storage (MinIO) as docker container

• forward the control plane and API Gateways ports locally

• create a dedicated LangStream CLI profile to interact with the cluster

• wrap all the common k8s tools to inspect the cluster (mini-langstream kubectl, mini-langstream helm, mini-langstream k9s)

1. Start or ensure the cluster is running:

mini-langstream start

2. Try to use the CLI:

mini-langstream cli apps list

3. Deploy an application:

export OPENAI_API_KEY=<your-openai-api-key>
mini-langstream cli apps deploy my-app -app https://github.com/LangStream/langstream/tree/main/examples/applications/openai-completions -s https://github.com/LangStream/langstream/blob/main/examples/secrets/secrets.yaml

To delete all the storage and stop the cluster:

mini-langstream delete

Your first application

Here are a few ways to get started building LangStream applications:

• Build a sample application to quickly create an OpenAI query with LangStream.

• Set up your development environment and learn how to build the needed manifests

• Install the VSCode extension and use the provided starter applications & agent snippets