Skip to main content

​
Overview

Knowledge Stack uses GitHub Actions for continuous integration and deployment. The pipeline is split into three workflows that automate testing, building, and publishing.

​
Workflows

​
Test Workflow

Runs automatically on every pull request and push to the main branch. It executes three jobs in parallel:
  • Integration tests — Runs the full test suite with isolated database containers
  • End-to-end tests — Spins up the complete stack and validates full user workflows using the generated SDK
  • Linting and type checking — Ensures code quality standards are met
When all checks pass on the main branch, a release job runs automatically to determine if a new version should be published.

​
Build and Deploy Workflow

Triggered automatically when a new version is released. This workflow:
  1. Builds production Docker images for the API and worker services
  2. Pushes images to the container registry
  3. Deploys updated images to the Kubernetes cluster
  4. Runs database migrations

​
Publish Workflow

Also triggered automatically on new releases. It publishes the generated SDKs:
  • Python SDK — Published to PyPI
  • TypeScript SDK — Published to npm
Both SDKs are generated from the OpenAPI specification, ensuring they always match the latest API.

​
Versioning

Knowledge Stack uses semantic versioning with automatic version bumps based on commit messages: Release tags follow the format v1.24.0.

​
Self-Hosted Deployments

If you are running your own Knowledge Stack instance, you can use the same CI/CD patterns:

​
Building Docker Images

​
Running the Test Suite

​
SDK Generation

Generate client SDKs from your running instance: