Concepts and getting around

Projects and workspaces

Once you sign up on Laminar, you will get to your dashboard. Here you can see and manage your workspaces and projects. When you create an account, you are created your personal workspace by default.

A Laminar project is a collection of pipelines, endpoints, datasets, evaluations, and execution logs. Access API keys and environment variables are configured at the project level as well. Think about it as a scope of a certain work.

A Laminar workspace is a collection of projects accessible and users that have access to them.

On pro and start-up tiers, you can create more workspaces and add people to them. This enables collaboration including real-time editing of the pipelines. To do that, you only need to know the email address with which they have registered to Laminar.

  1. Navigate to the projects page by clicking the Laminar’s logo in the top left corner.
  2. Navigate to the workspace by clicking on the workspace name on the left.
  3. Click “Add user”.
  4. Type in their email and click “Add”.

In the same page, you can create projects.

Pipelines

Once you create your first project, you will be redirected to the pipelines page. Pipeline builder is the place where you build your core logic, evaluations logic, and experiment with the flows. For more information on pipelines and navigating in the builder page, see the pipelines documentation.

Endpoints

Think of every pipeline that you build as a large function with input parameters and one or many outputs. Laminar allows you to deploy every pipeline to endpoints and then call it externally from your code.

Every input node of the pipeline will serve as an input parameter in the API call and the outputs will be returned in the response. See endpoints documentation for more.

Evaluations and datasets

Laminar has a powerful set of tools for evaluations where you can create any custom pipeline to evaluate how your other pipelines work or evaluate the data in your datasets. In order to run evaluations, you need to create a dataset, add data to it, and create an evaluator pipeline. See more in evaluation documentation.

Evaluations run in parallel batches, and you can get the results faster than you ever had seen it before.

Onboarding to Laminar

Read onto the following sections to learn how to migrate your previous workflows to Laminar or start from scratch.