duck-ui
The fully open-source DuckDB workbench that runs in your browser. SQL editor, notebooks, charts, AI assistant. No install, no signup, no backend — your data nev
What is duck-ui?
The fully open-source DuckDB workbench that runs in your browser. SQL editor, notebooks, charts, AI assistant. No install, no signup, no backend — your data never leaves the tab.
Overview of duck-ui
What it is
Duck-UI is a fully open-source DuckDB workbench that runs entirely in the browser. It requires no installation, no signup, and no backend server. All data processing stays within the browser tab; data never leaves the tab. The application is built with Svelte 5 and CodeMirror 6, with a startup bundle of about 350 kB of JavaScript.
Key features
- Workspace with pages and tabs for queries, notebooks, and dashboards; quick navigation via ⌘K.
- SQL editor with CodeMirror, completion from catalog and DuckDB, DuckDB dialect formatting, and EXPLAIN ANALYZE viewer.
- Virtualized results grid with sort, filter, search, cell selection, column statistics, and export to CSV, JSON, XLSX, and Parquet.
- Notebooks supporting SQL, Python (via Pyodide with pandas DataFrames), and markdown cells with individual results and charts.
- Dashboards as markdown reports with named SQL blocks, charts, tables, inputs, and auto refresh; shareable via link.
- Duck Brain AI assistant for text-to-SQL and error fixing, supporting WebLLM (fully local, no API key), OpenAI, Anthropic, or any OpenAI-compatible endpoint.
- Import data from CSV, JSON, Parquet, Arrow, XLSX, .duckdb files, or remote sources (S3, GCS, Azure, R2, MinIO); local folders via File System Access API.
Use cases
- Ad-hoc data analysis and SQL querying directly in the browser without any setup.
- Building and sharing interactive notebooks and dashboards with SQL, Python, and markdown.
- AI-assisted query generation and error fixing using local or cloud-based language models.
- Embedding live, runnable queries in web pages via iframe without a server.
- Publishing read-only datasets with kiosk mode for zero-backend data sharing.
- Offline and air-gapped environments where no external server or internet access is available.
Getting started
Run the Docker image:bashdocker run -p 5522:5522 ghcr.io/caioricciuti/duck-ui:latestThen open http://localhost:5522 in your browser. No further setup is required.
License and pricing
The project is licensed under MIT and is open source. No pricing information is provided; it is free to use.
Based on the project's public README and repository metadata.
