LangStream AI extension for VSCode

The LangStream project combines the intelligence of large language models (LLMs) with the agility of streaming processing, to create powerful processing applications. An application in LangStream can watch a message topic and process data through multiple steps to output some useful generative AI results. Learn more about the LangStream project here. Or get started with the LangStream documentation here.

This extension provides ways to interact with the LangStream API from within VSCode. It provides a simple way to create, update, and delete LangStream applications, modules, and pipeline agents. It also provides a way to view the status of your applications and message with its gateway.

For more about LangStream visit the project website.

Features

• Environment
  • Save and load control plane configurations
  • List control plane tenants/applications
  • API client
• Tenant
  • Create and delete tenants
  • List tenant applications
  • Initialize application manifests
• Application
  • List application components (modules/pipelines/agents)
  • Delete application
  • Get application details
  • Watch and search logs
  • Message with gateways
• Gateway
  • List gateways
• Pipeline
  • Code lens offers a "deploy" and "update" button to deploy or update a pipeline
• Agent
  • Snippet completion while building a pipeline

Starting a LangStream environment

With kubectl and helm installed, you can start a LangStream control plane with the following commands.

kubectl apply -f https://raw.githubusercontent.com/LangStream/langstream/main/helm/examples/minio-dev.yaml
helm repo add langstream https://datastax.github.io/langstream
helm repo update
helm upgrade \
    -i langstream \
    -n langstream \
    --create-namespace \
    --wait \
    --values https://raw.githubusercontent.com/LangStream/langstream/main/helm/examples/simple.yaml \
    langstream/langstream

Forward the control plane and gateway ports

# Control plane
kubectl port-forward -n langstream svc/langstream-control-plane 8090:8090

# Api gateway
kubectl port-forward -n langstream svc/langstream-api-gateway 8091:8091

Saving the environment in the extension

Once the extension is installed, the activity bar will have a new LangStream icon. Click it to activate the extension. You will see a link to "Add a control plane". Click that link to show the wizard. It will ask for:

• Control plane address (default http://localhost:8090)
• API Gateway address (default http://localhost:8091)
• Name the gateway

With the environment saved all its objects will be discovered and listed.