Quick Start Guide

Create an LLM pipeline using one of our ready-made templates and enjoy the benefits of deepset AI Platform.

In a nutshell, here are the steps to create a search app:

Set Up Your Workspace

In deepset AI Platform, you work in workspaces, where you store your files, indexes, and pipelines. You can have up to 100 workspaces in your organization. Let's start by creating a workspace dedicated for your data:

  1. Log in to deepset AI Platform.
  2. Click your initials in the upper right corner and choose Workspaces.
  3. Click Add workspace, give your workspace a name, and save it.

Upload Your Files

📂

Sample Files

You can also use one of the sample datasets we prepared. They're available on the Files page.

Add the files on which you want to run your search:

  1. Make sure you're in the workspace you just created, and in the left navigation, go to Files > Upload Files.
  2. Drag the files from your computer and drop them into the Upload Files window.
  3. Click Upload. Your files are now listed on the Files page.

ℹ️

Other Upload Options

You can also use deepset SDK or REST API endpoints. For details, see Upload Files.

Create an Index

Indexes define how your files are prepared for search and specify the document store where they’re saved. For an easy start, use a template:

  1. Go to Indexes and click Create Index.

  2. Choose the Standard Index (English) template. This template works for most use cases.

  3. Click the template card, give your index a name, and click Create Index.

  4. Click More actions next to your index and choose Enable. Your files are being indexed. Once the indexing is complete, your query pipeline can run searches on the indexed files.

ℹ️

Indexes

You can reuse indexes among your query pipelines. To learn more about indexes, see Indexes.

Create a Pipeline

Pipelines in deepset AI Platform, are the engines powering your apps. A pipeline defines the steps your app takes to resolve a user query.

Now, create a pipeline in the simplest way:

  1. In deepset AI Platform, go to Pipeline Templates.

  2. Find a template that best matches your use case, hover over it, and click Use Template.

    A highlighted template of a RAG pipeline that detects prompt injection with the Use Template button highlighted.
  3. Give your pipeline a name and click Create Pipeline. You're redirected to the Pipeline Builder.

  4. In your pipeline components, find OpenSearchDocumentStore and choose the index you previously created. It's shown in the list on the component card.

    Document store component card
  5. Save your pipeline.

  6. To use your pipeline, you must deploy it first. Click Deploy in the upper right corner.

    Pipeline details with the Deploy button highglighted

ℹ️

More about pipelines

For a deeper dive into pipelines and components, see Pipelines and Pipeline Components.

Test Your Pipeline

  1. In deepset AI Platform, click Playground.
  2. Select the pipeline that you just deployed.
  3. Type a query and click Search to try out your pipeline. That's it!

What To Do Next?

Have a look at our tutorials, we suggest you start with the basic ones available at Learn the Basics.

For guides on how to perform tasks in deepset, see the How-to Guides section.

To learn about how things work in deepset and understand the concepts, see the Concepts section.