DuckDB online: run DuckDB in your browser with no install

Five tools run DuckDB SQL from a browser tab. Three run DuckDB inside the tab with DuckDB-WASM, MotherDuck runs it in the cloud, and the DuckDB UI runs it on your own computer.

DuckDB Web Shell

The official DuckDB Web Shell at shell.duckdb.org gives you a DuckDB prompt in a browser tab, running on DuckDB-WASM. It reads files from HTTPS URLs, and the .files add command opens a picker for local files. It is free with no account, and it works like a terminal, so results print as text rather than in a grid.

rows.page

rows.page puts a scrollable grid, a filter box and a Columns panel with histograms on top of DuckDB-WASM. You drop a CSV, TSV, JSON, JSONL or Parquet file on the page, or paste cells copied from Excel or Google Sheets, and the data stays in your browser unless you press Share. It is free with no signup, and the Pro ($12 per month) and Max ($39 per month) plans are for hosting shared datasets.

SQLite Online

SQLite Online runs SQLite, DuckDB and PGlite in your browser with no signup, so you can try one query on three engines. You import a CSV into a table first, then query it.

MotherDuck

MotherDuck is a cloud DuckDB service, so tables you load stay in your account between sessions. In its web UI, the Add data button uploads a CSV, JSON or Parquet file. The Lite plan is free, and Business starts at $250 per organization per month plus usage.

DuckDB UI on your computer

Since version 1.2.1 (March 2025), the DuckDB CLI can open a notebook UI built by MotherDuck. It runs DuckDB as a native process, so queries use all of your computer's cores, memory and files, and by default your data never leaves your machine. It is free, but it does not support Intel Macs or Windows on ARM.

1. Install the DuckDB CLI, version 1.2.1 or newer.
2. Run duckdb -ui in a terminal.
3. The UI opens in your browser at http://localhost:4213.

1. Open rows.page and drop a CSV, TSV, JSON, JSONL or Parquet file on the page, or click Paste data.
2. Type a quick filter such as country:NL or price:>20 in the filter box.
3. For SQL, type one statement that starts with SELECT, WITH or FROM in the same box, using data as the table name.
4. Export the result as CSV, Parquet, JSON, JSONL or Markdown, or press Share to get a link.

SELECT
  "user.address.city" AS city,
  count(*) AS orders,
  median(amount) AS median_amount
FROM data
WHERE status = 'paid'
GROUP BY ALL
HAVING count(*) > 10
ORDER BY orders DESC

Window functions, QUALIFY, unnest(), PIVOT inside a subquery and FROM (SUMMARIZE data) all work, and results show up to 1,000,000 rows. Nested JSON becomes dot-path columns up to 8 levels deep, so quote names with dots, like "user.address.city". A hidden __rid column holds the original row number, so remove duplicates with SELECT DISTINCT * EXCLUDE (__rid) FROM data.

DuckDB-WASM is DuckDB compiled to WebAssembly, so queries run inside your browser tab and your file is not uploaded. A tab can use at most 4 GB of WebAssembly memory. rows.page adds these limits:

• CSV, TSV, JSON and JSONL files open up to 1.5 GB per tab.
• A bigger CSV, TSV or JSONL file loads its first rows, with the note "Loaded the first N rows: the file is X, more than the browser can hold."
• A .json file over 1.5 GB is refused with "Too big for this browser", so convert it to JSONL to open part of it.
• Parquet skips the 1.5 GB check, because DuckDB reads it in place.
• One SQL statement on one file: no cell editing, no CREATE TABLE, and no joins across two files.
• SQL results and their exports stop at 1,000,000 rows, while a filtered view exports every matching row.
• No Excel, SQLite, XML or Avro files, so save a sheet as CSV first.

• A quick query on a file you have, or a link to share it: rows.page.
• A plain DuckDB prompt for scripts with several statements: the DuckDB Web Shell.
• The same query on DuckDB, SQLite and PostgreSQL (PGlite): SQLite Online.
• Tables you want to keep and query again from any device: MotherDuck.
• Files bigger than a browser tab can hold, kept on your own machine: the DuckDB UI.