Browse Iceberg Tables Online — No Spark, No Trino

Apache Iceberg tables are everywhere now — but browsing them still feels like it requires a PhD in distributed systems. PondPilot lets you explore Iceberg tables from a browser tab. No Spark session, no Trino cluster, no Docker compose files.

Connect and Browse

PondPilot connects to Iceberg REST Catalogs using DuckDB’s Iceberg extension running in WASM. Once connected:

  • Namespaces — see all databases/schemas in your catalog
  • Tables — list tables within each namespace
  • Schemas — inspect column names, types, and nullability
  • Data — preview rows or run full SQL queries

No Setup Required

The traditional path to querying an Iceberg table:

  1. Install Java
  2. Set up Spark or Trino
  3. Configure catalog properties
  4. Write a query
  5. Wait for the session to start

The PondPilot path:

  1. Open app.pondpilot.io
  2. Enter catalog URL
  3. Query

That’s a meaningful difference when you just want to check if yesterday’s pipeline ran correctly.

SQL Queries on Iceberg Data

Full DuckDB SQL works against Iceberg tables:

SELECT date_trunc('hour', created_at) as hour,
       COUNT(*) as events
FROM catalog.default.clickstream
WHERE created_at >= '2025-01-01'
GROUP BY 1
ORDER BY 1

Who This Is For

  • Data engineers validating pipeline outputs
  • Analytics engineers exploring new tables before building models
  • Data scientists who want to look at the data before writing code
  • Platform teams who need a lightweight catalog browser for their org

Limitations

This is browser-based compute. It’s great for exploration, validation, and small-to-medium queries. For scanning terabytes of data, you’ll want server-side DuckDB or a proper compute engine. But for the vast majority of “browse and check” workflows, it’s more than enough.

Get Started

Open app.pondpilot.io and connect your Iceberg catalog.

Back to home