LakeDB
GitHub
0.9.0 Clearer QuerIA

Ask. Clarify.
Run when ready.

Prepare SQL in clear query cards. Refine a selected query, resolve real ambiguities and choose exactly what to run. Initial Redshift support is available for beta testing; live AWS validation is pending.

Download 0.9.0 What’s new

Free local SQL client macOS, Windows & Linux

LakeDB 0.9.0 presents ask, clarify and run when ready
0.9.0 Ask. Clarify. Run when ready.
One local workspace.MySQLMariaDBPostgreSQLSQLite

01 / What’s new in 0.9.0

Keep each query clear.
Choose your next step.

Separate SQL cards, focused clarification and explicit execution make QuerIA easier to follow.

01

A card for each query

Each request owns its SQL and numbered Run action. Review it before executing; generating SQL does not run it.

Explicit execution
02

Refine or ask independently

Refine a selected query with its previous question and SQL. Focused clarification resolves real ambiguity after metadata inspection. Rows and results stay local.

Clearer context
03

Simpler Enterprise access

Login connects to your assigned organization. Pending requests keep a Check access action so you can continue after approval without starting over.

You choose when to leave
04

Initial Redshift support

Choose the official AWS Python connector or AWS Data API. Live AWS validation is pending before stable 1.0. Unsupported tools remain disabled.

Beta testing
Read the 0.9.0 release notes

02 / Compare & migrate

Understand the difference.
Then decide what to move.

A model check, a data check and a migration answer different questions. LakeDB keeps those choices visible.

  1. 1

    Choose the scope

    Select source, target and objects. Begin with model comparison; enable data checks only when they answer your question.

  2. 2

    Inspect the differences

    Review structures, object definitions and the checks you requested. Export the generated SQL or prepare a migration plan.

  3. 3

    Review and run

    Check the target, dependencies and copy options before executing through your own local database connection.

Table analysis opens with Model selected; data checks are opt-in. Illustrated with demo schemas. Click to inspect.
Choose the check that answers your question
CheckWhat it tells youDatabase work
Model DefaultWhether the selected structures and definitions match.Metadata inspection; no row-content comparison.
Exact row countsWhether selected tables contain the same number of rows.Can scan table indexes. Matching counts do not prove matching values.
Content / checksumsWhether the selected content checksums match.Reads table data and can be expensive on large tables.
Merge

Keep target-only rows.

Reads all source rows and inserts or updates target rows. It is a complete source pass, rather than a row-by-row difference download.

Replace data

Clear, then copy.

Deletes target rows before copying the source. Dependent tables may be required. Review the plan and the target before execution.

On a busy production database, start with model comparison and a small scope. Choose exact counts, content checks and data copies deliberately; their work grows with the selected tables. DDL rollback depends on the engine.

03 / Your everyday SQL client

Each connection
keeps its own context.

SQL tabs, table tabs, selected schema and layout belong to their connection. Move between local, staging and production without mixing workspaces.

SQL editor, object explorer and results in one connection workspace. Click to inspect.

Write with context

Schema-aware completion, local syntax diagnostics, formatting, Explain, history and customizable shortcuts.

Edit with review

Typed staged edits, stable row identities, conflict checks, transactions and streaming exports.

Connect your way

Direct, SSH and TLS routes for server engines. Local SQLite files, read-only connections and production confirmations.

Native behavior, explicit capabilities

MySQL & MariaDBServer catalogs, access management, backup, restore, comparison and migration.

PostgreSQLNative schemas, types and plans, access management, schema backup/restore and transactional table copy.

SQLiteLocal file connections, SQL, browsing, table editing and exports. Unsupported transfer tools stay disabled.

04 / Optional QuerIA

Ask in natural language.
Review in SQL.

QuerIA sits beside your SQL tabs. Draft a focused query with Normal generation, or inspect relationships and indexes with Agentic generation. The complete SQL stays visible before you run it.

Generated SQL is a proposal. You choose whether to execute it.

Starts disabled

Review the privacy notice and activate the optional service explicitly. You can disable it again.

Rows stay local

Database credentials, table rows and query results are never sent to the AI service.

Limited, temporary context

Requests use your question, chosen context and relevant schema metadata. LakeDB Service does not retain those inputs or generated SQL.

Read the privacy policy

The complete local SQL client is free to use. Optional Patreon membership supports development and provides higher QuerIA usage limits.

05 / Road to 1.0

Useful today.
Getting ready for 1.0.

0.9.0 improves QuerIA and adds initial Redshift transports. Live AWS tests, compatibility checks and trusted distribution remain gates before stable 1.0.

Product stages and quality gates, rather than promised release dates.
Read the complete roadmap

06 / Download 0.9.0

Ready for your next query.

Choose the package for your computer. All downloads come from the official GitHub release.

Release notes & SHA-256 files

Using a previous 1.0.0-beta.*, 0.10 or 0.11 build? Install 0.9.0 manually once; keep your existing application data. Older updaters consider this new numbering a lower version.

Prefer Homebrew?

Install the official cask on macOS Apple Silicon.

brew tap DavLagoHern/lakedb
brew trust DavLagoHern/lakedb
brew install --cask lakedb
Does updating change my databases?

Installing an update does not migrate your connected databases. Connections, preferences and workspaces remain in the local profile. Database changes only run when you explicitly execute SQL, apply edits or start an operation.

Do all engines have the same tools?

No. LakeDB exposes engine capabilities explicitly. MySQL/MariaDB and PostgreSQL have different transfer and catalog behavior; unsupported tools remain disabled for SQLite. See the documentation for each engine.

How can I report a problem?

Use the bug report form with your LakeDB version, operating system, engine version and steps to reproduce. Review diagnostics before sharing and remove sensitive data.

Independent development

Help shape the next update.

Share a real workflow, report a reproducible bug or support continued development.

Suggest an idea Support on Patreon