1. DBeard
  2. Databases
  3. DuckDB

DuckDB desktop GUIstable

Big questions. One DuckDB file.

Local analytics, with your schemas and SQL in one desktop window.

macOS first, with Linux to follow early. Windows is planned.

Local DuckDB file support is stable. A file held open for writing by another program must be released before DBeard can open it. MotherDuck is a separate, untested integration.

The DBeard pirate opens a treasure chest filled with charts and marked with the DuckDB logo.
Verified scope
Local DuckDB files
Connection type
Open a file
Read-only mode
Available in connection settings

The lay of the land

Everything DuckDB keeps, where DuckDB keeps it

DBeard lists each object at the level the engine scopes it, so the sidebar matches what you would see from the engine’s own catalog.

~/Projects/analytics.duckdb
  • Open or drop a local DuckDB file into DBeard. File contents identify it as DuckDB rather than SQLite.
  • Choose Read only for a connection to open the file without write access.
  • Release a file held open for writing by another program before opening it in DBeard.
  • Extensions required by a file must already be installed. Opening a file does not automatically download them; an explicit INSTALL statement in a query tab can.
  1. DuckDB file connection
  2. Schema folderShown when the file has objects in multiple schemas.
  3. Tables and viewsA single-schema file lists Tables and Views under the connection.
  4. Columns
  5. Object groupsExpand groups for sequences, macros and types, including enums.

Made for DuckDB

What you get on this engine

  1. More than a table list

    Browse schemas, tables, views, sequences, scalar and table macros, and enumerated types. Inspect table indexes and constraints alongside your data.

  2. Review the rows you change

    Stage edits and review their SQL before applying them. Rows use their primary key, or rowid when no primary key exists, with conflict checks before changes are committed.

  3. Open files with clear boundaries

    Choose Read only when you need to inspect a file. File-lock errors identify the program holding it when DuckDB provides that information. Opening a file never downloads extensions automatically.

Demo coming soon
Inspect a local DuckDB file with its schema and macros in view.

Support

DuckDB features, one by one

“Coming soon” means it is on the roadmap. “Not planned” means we have decided against it.

Local workflows
FeatureDBeard
Detect and open DuckDB filesYes
Read-only connectionsYes
Browse schemas, sequences, macros and enumsYes
Inspect indexes and constraints in table structureYes
Stage grid edits and review SQLYes
MotherDuck connectionsRequires a separately installed compatible signed extension and token. Live verification is pending.Partial

DuckDB benchmarks

Benchmark results coming soon.

Ports of call

Hosted DuckDB it connects to

ServiceStatusNotes
Local DuckDB fileTestedThe repo records tests against local database files.
MotherDuckShould workRequires a separately installed compatible signed extension and a token. No live verification.

On deck

DuckDB in DBeard, up close

SELECT schema_name, function_name
FROM duckdb_functions()
WHERE function_type = 'table_macro'
ORDER BY schema_name, function_name;

The longer read

A DuckDB client worth switching to

Open the file. Keep the context.

A local DuckDB database belongs in your analytics workflow without another server to configure. Open the file, explore schemas and inspect tables and views. Macros, sequences and enum types appear with the schema they belong to, so you can understand how a dataset was put together before writing your next query.

Use a query tab for SQL and the table browser for inspecting rows. DBeard is a native Rust and GPUI app on your Mac.

Make deliberate changes

Rows are addressed by their primary key, with rowid used for tables without one. Review staged edits and their generated SQL before applying them. Conflict checks refuse changes when a row has changed underneath you, and grouped edits apply together in a transaction.

Choose Read only when the job is inspection. A transaction started with BEGIN in a query tab makes grid edits wait; Manual commit uses a separate connection. Keep that distinction in mind when working across query tabs and the grid.

Understand locks and extensions

If another program has the file open for writing, release that connection before opening it in DBeard. The app surfaces the file-lock error and the holder’s name when supplied by DuckDB. See DuckDB’s official concurrency documentation for its file access model.

Opening a database never automatically downloads an extension. An installed extension can be loaded when a file needs it; a missing extension is reported. Explicitly running INSTALL in a query tab remains possible.

Local DuckDB and MotherDuck have different scopes

This page describes local DuckDB files. MotherDuck accepts md: database and share addresses, requires token authentication and a compatible signed extension installed separately, and remains untested against the real service. Do not carry the stable local-file label over to cloud access.

Questions

DuckDB and DBeard, answered

Can I open an existing DuckDB file?

Yes. Open or drop the file into DBeard. The app detects DuckDB from its contents and lists its schemas, tables, views, macros, sequences and enum types.

Why can I not open a file that another program is using?

A file held open for writing by another program must be released first. DBeard reports the lock and identifies the program when DuckDB includes that information. Close the writing connection and then reopen the file.

Does opening a file install DuckDB extensions?

No. Opening a file uses required extensions that are already installed and reports missing ones. It does not download extensions automatically. You can explicitly run INSTALL from a query tab.

Can I edit rows while a transaction is open?

A transaction started with BEGIN in a query tab makes grid edits wait because DuckDB has no savepoints. A query tab using Manual commit has its own connection, so grid edits do not wait for that transaction.

Is MotherDuck covered by stable DuckDB support?

No. MotherDuck is a separate, untested integration. It requires a separately installed compatible signed DuckDB extension and token authentication. Local DuckDB file testing does not verify MotherDuck service access.

Open your DuckDB database in DBeard.

Free for one connection at a time, with every feature. No account, and your data stays on your machine.