Automation & Workflow Options
Extending ĢƵ to External Workflows
Every engineering team works a little different. In this lesson, we’ll share a few options for extending ĢƵ to external workflows. This is helpful if you want to automate processes, design experiments as a quality gate in your CI/CD process, or work with LLMs.
The four options we’ll cover are the ĢƵ API, CLI, GitHub Actions, and the ĢƵ MCP Server.
Using the ĢƵ API
With the , you can make requests to the platform directly to execute a wide range of actions. Our API is written according to the OpenAPI 3.0 Specification and you can review here to see all the types of requests you can make.
To get started, go to Settings → Access Tokens and create a new token you can use to authenticate your API calls.
With a Team Token, you can make requests on experiments related to that Team.
With an Admin Token, you can make requests to update teams and environments.
If you want to get into more details, you can review our full guidance on using the ĢƵ API in .
Using the ĢƵ Command Line Interface (CLI)
If you prefer, you can also use the ĢƵ CLI to do similar actions. You’ll need to install a simple NPM package with the following command:
npm install -g steadybit
Then, you’ll create a profile with your access token for authentication.
➜ steadybit config profile add
? Profile name: steadybit
? API access token: [hidden]
? Base URL of the ĢƵ server: https://platform.steadybit.com
If you need more details, you can reference our .
Using ĢƵ with GitHub Actions
If you want to build GitOps workflows with ĢƵ, you can also use a GitHub Action to initiate experiments. In the GitHub Action marketplace, you can find the پDz.
For more information, you can .
Using the ĢƵ MCP Server with LLMs
More and more teams are incorporating AI and LLMs into their daily work. The ĢƵ allows you to easily connect data and functionality with your LLM of choice so you can run new reliability workflows.
For example, you could use an LLM to analyze your recent experiment runs to surface trends and make data-based recommendations about what experiments to run next. ĢƵ extends easily so as you come up with a new idea or prompt, you can test it out fast and see value.
Lesson Summary
Now you have a high-level view of all the ways you can integrate ĢƵ into your external workflows to automate experiment runs, experiment creation, get custom reports, and more.
With the ĢƵ API, CLI, MCP Server, and GitHub Actions; you can seamlessly incorporate reliability into any development lifecycle.

