Skip to main content

Pipeline Quick Start Guide

Overview​

This guide walks you through the essential steps to create, configure, and run your first data pipeline in Rhombus AI. Follow these instructions to get started quickly and efficiently.

Before You Begin​

Before creating your first project, please sign in to your Rhombus AI account:

  • Go to www.rhombusai.com and click Sign In.
  • If you don’t have an account, click Sign Up to create one.

After signing in, you’ll be automatically redirected to the Workflow page where you can start creating your first project.

Step 1 — Create Your First Project​

You can create a new project in two different ways:

Option A — From the Left Panel​

  1. Open the Projects panel on the left.
  2. Click New Project.

Option B — From the Main Area​

  1. In the center of the page, click Create Project.

Finalize Project Creation​

  1. Enter a project name.
  2. Click Create.
  3. Your new project will appear in the Project List on the left panel.
tip
  1. A project acts as a workspace to store your datasets, transformations, and results.
  2. Free plan users can create up to 3 projects. To unlock unlimited projects, upgrade your plan.

Create Project

Step 2 — Upload Your Dataset​

  1. In the Chatbot panel, click the “+” button.
  2. Select Add New File to upload from your local computer, or choose Third-Party Data (e.g., AWS S3) to import from supported cloud storage.
  3. Once the upload or import is complete, the file will appear in the Chatbot panel and can be used in your pipeline.
tip
  1. Supported file formats include CSV and XLSX (Excel).
  2. Free plan users can upload files up to 10 MB per file, with a maximum total size of 50 MB across all files.

Upload

Step 3 — Chat with the AI Agent​

  1. After your dataset has been uploaded, enter your request in the Chatbot panel.
  2. Type your request, for example:
    “Clean and impute the dataset.”
  3. The AI agent will suggest or automatically create the required transformation steps for you.
tip
  1. The AI agent helps reduce manual setup by translating your natural language instructions into pipeline actions.

Chabot

Step 4.1 — Preview Your Dataset​

  1. Click on the last node in your pipeline to inspect the final output of your transformations.
  2. You can preview the processed dataset in two different ways:
    • Option A — Preview Tab
      Open the Preview tab to view sample rows of your dataset.
    • Option B — Quick Preview
      Click the eye icon in the top-left corner of the node to quickly preview the dataset.
  3. Review the previewed data to confirm whether the output matches your expectations.
tip
  1. The preview shows only a subset of your dataset for faster loading and smoother performance.
  2. To view the full dataset, you can download the processed results from the Export section.

preview

Step 4.2 — Refine Your Pipeline (Optional)​

If the preview in Step 4.1 shows unexpected results or data quality issues, you have two ways to refine your pipeline:

Option A — Use the AI Agent​

  1. Continue chatting with the AI agent in the Chatbot.
  2. Provide additional context or instructions.
  3. The agent will automatically update your pipeline based on your input.

Option B — Manually Adjust the Pipeline​

  1. Double-click an existing transformation node to tweak its parameters.
  2. Drag and drop additional transformation nodes from the Palette if needed.
  3. Recheck the last node preview to verify your updates.

Additional Dataset Inspection (Optional)​

  • Open the Dataset Profile in the side panel.
  • Review the schema, field types, and statistics for deeper inspection.
tip
  1. If you encounter an error message, click “Fix in Chat” to let the AI agent assist you in resolving the issue.

Configuration

Step 5 — Run & Download Your Dataset​

After you’ve refined your pipeline, you need to run it before downloading the processed dataset.

Option A — Automatic Run (via Chatbot)​

  • If you made changes through the Chatbot, the pipeline will run automatically.
  • Wait until the run completes, then proceed to download your dataset.

Option B — Manual Run (via Apply Button)​

  1. If you manually adjusted nodes or added new transformations, click the last node in your pipeline.
  2. Click Apply to run the pipeline manually.
  3. Once the run is complete, confirm the processed data in the Preview tab.

Download Your Dataset​

  1. In the last node’s Preview tab, click Download Dataset.
  2. Choose your preferred format:
    • CSV
    • Excel (XLSX)
tip
  1. The downloaded file contains the final processed dataset based on your configured transformations.
  2. If the output is not as expected, revisit Step 4.2 to refine your pipeline and re-run it.

download