DuckLocal

Analyze local data files with DuckDB SQL

Open CSV, Parquet, JSON, Excel and DuckDB files directly in a desktop workspace. Explore the schema, write SQL, chart the answer and export the result—without uploading your data or setting up a server.

Download DuckLocal Try the 10-minute tutorial

Free and open source for macOS, Windows and Linux. No account. No telemetry.

From local files to useful results

Skip the upload, import wizard and database setup. Start with the files you already have.

  1. Open files, folders or patterns

    Drop in one file, open a directory recursively, or use a pattern such as ./data/2026-0*/*.parquet. DuckLocal registers each supported file in the workspace.

  2. Inspect before you query

    See relations, columns and inferred types in the schema sidebar. CSV, Parquet and JSON files become views; every spreadsheet sheet becomes a table.

  3. Query, chart and export

    Run DuckDB SQL in the editor, switch a result to a chart, inspect each column, or export the complete result to CSV or Parquet.

Ask the question where the data lives

DuckDB scans the source file locally. Aggregations happen in SQL, so a multi-million-row file can produce a small, useful answer without first loading every row into a browser or sending it to a service.

Explore the SQL editor · See results and charts

SELECT
  channel,
  round(sum(revenue), 2) AS revenue
FROM sales
GROUP BY channel
ORDER BY revenue DESC;

One workspace for the formats you already use

Work across individual files, whole folders and databases without converting everything first.

  • .csv .tsvTypes inferred with DuckDB
  • .parquetFast columnar scans and metadata
  • .json .jsonlDocuments and newline-delimited records
  • .xlsx .odsEvery non-empty sheet becomes a table
  • .duckdbOpen a database or attach it beside another
  • folder/Recursively find supported data files
  • *.parquetOpen matching files with glob patterns
  • s3://Connect only when you explicitly choose to

Read the complete data-source guide →

Your analysis does not need an upload step

DuckLocal is a native desktop application with DuckDB embedded. Local files are read from their existing paths and are not copied to a DuckLocal cloud service.

Network features are explicit: S3 queries, map tiles and update checks only apply when you choose to use them.

  • No account or workspace provisioning
  • No telemetry and no DuckLocal cloud service
  • No mandatory file conversion or database import
  • Source files remain untouched when a view is removed
  • Open-source code under Apache-2.0

More than a query box

Keep the schema, SQL editor, results, chart, query history and profiler together. Attach DuckDB or SQLite databases, inspect column distributions and come back to registered files on the next launch.

Local data analysis questions

Can DuckLocal analyze files without uploading them?
Yes. DuckLocal reads local files where they are and runs DuckDB on your computer. It does not upload those files to a DuckLocal cloud service, and it requires no account.
Which local file formats can DuckLocal query?
DuckLocal opens CSV, TSV, Parquet, JSON, JSON Lines, Excel and OpenDocument spreadsheets, DuckDB database files, folders and glob patterns. It can also connect to S3 when you explicitly configure it.
Do I need to create a database or schema first?
No. Data files become queryable views in DuckDB, while spreadsheet sheets become tables. DuckLocal infers columns and types and shows them in the schema sidebar.
Can I export the complete query result?
Yes. DuckLocal can export a query result to CSV or Parquet. Export reruns the query and writes the complete result rather than only the rows visible in the grid.

Run your first local query

Download DuckLocal and use the tutorial’s sample CSV to go from file to chart in about ten minutes.

Download DuckLocal Open the tutorial