Hyperparam Quick Start Guide

Get started with Hyperparam in just 3 minutes! This guide will walk you through your first dataset exploration.

Step 1: Access Hyperparam

Navigate to hyperparam.app in your web browser.

You'll see the home page with a drag-and-drop area:

Hyperparam Home

Step 2: Load Your First Dataset

You have three options to load data:

Option A: Drag and Drop a Local File

  • Drag any .parquet file directly onto the drop zone
  • The file loads instantly in your browser (no upload required!)

Option B: Load from URL

  • Click on one of the example files, or
  • Paste any public URL to a Parquet file

Option C: Discover Datasets via Chat (Requires Sign-in)

  • Click "Hyperparam Chat"
  • Sign in with Google
  • Ask for datasets, e.g., "Find me a conversation dataset with at least 100k examples"

Step 3: Explore Your Data

Once loaded, you'll see the interactive table view:

  • Scroll infinitely through millions of rows
  • Click column headers to sort
  • Double-click cells to inspect nested data
  • Drag columns to reorder

Key Navigation Tips:

  • Use mouse wheel or trackpad for smooth scrolling
  • Click any cell to see its full content
  • For JSON/structured columns, double-click to expand

Step 4: Unlock AI Features (Sign In)

To access AI-powered features, sign in with Google.

Once signed in, you can:

  • Use chat to find and analyze datasets
  • Add AI-generated quality scores
  • Create workspaces for editing
  • Export curated datasets

Step 5: Try AI-Powered Analysis

With sign-in complete, try these powerful features:

In the Main Chat:

"Find me a chat dataset that includes reasoning examples"

In the Table Chat:

"Add a column that rates response quality from 0-10"

Your First Workflow

Here's a simple workflow to try:

  1. Load a dataset (try the Wikipedia example)
  2. Sign in to unlock AI features
  3. Open table chat and add a quality column
  4. Create a workspace to work with a subset
  5. Sort by quality to find the best/worst examples
  6. Export your curated dataset

What's Next?

Need Help?

Welcome to the future of dataset engineering!

Hyperparam Quick Start Guide - Hyperparam