caioricciuti/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.

What it solves

Duck-UI provides a fully browser-based, serverless workbench for DuckDB. It solves the problem of needing to install software or manage backend infrastructure to analyze data, allowing users to perform SQL queries, run Python notebooks, and create dashboards entirely within their browser tab.

How it works

It runs DuckDB via WebAssembly (WASM) directly in your browser. Because it is client-side, your data never leaves your machine. It supports importing various file formats (CSV, Parquet, JSON, etc.) and connecting to remote data sources. It includes a "Duck Brain" AI assistant that can generate SQL or fix errors using either local WebLLM models (which run entirely offline) or external API providers like OpenAI or Anthropic.

Who it’s for

It is designed for data analysts, developers, and researchers who want a portable, privacy-focused tool to explore datasets, share analyses via URLs, or embed interactive SQL explorers into their own websites without managing a database server.

Highlights

  • Fully client-side execution with no backend or sign-up required.
  • Integrated AI assistant for Text-to-SQL and error fixing, with support for offline local models.
  • Supports SQL, Python (Pyodide), and Markdown notebooks with interactive charts.
  • Enables live collaborative sessions via WebRTC without a central server.
  • Portable deployment options including Docker, static hosting, and GitHub Pages.
  • Kiosk mode for publishing read-only, interactive data dashboards.

Related

  • Project
    geek-fun/dockit

    DocKit is a cross‑platform native desktop client (Tauri + Vue) for managing Elasticsearch, OpenSearch, DynamoDB, MongoDB and EasySearch. It offers Monaco‑based query editors, visual tools, import/export, and—behind a paid “Ultimate” subscription—a natural‑language AI agent that writes and runs queries using OpenAI, Anthropic, DeepSeek, Ollama, etc. The free Community edition provides full database connectivity with encrypted local credential storage and no telemetry.

  • Project
    motherduckdb/mcp-server-motherduck

    A self‑hosted MCP server that wraps DuckDB (local, S3, or MotherDuck) so AI assistants can run SQL queries via the Model‑Chat‑Protocol. Supports stdio or HTTP, read/write control, Docker deployment, and integrates with Claude, Gemini, Cursor, etc.

  • Project
    AnotiaWang/deep-research-web-ui

    Deep Research Web UI is an open‑source web app that turns a research question into a cited report. It combines LLMs with web‑search APIs, shows a tree‑view of the search process, lets you inspect and update each citation, and can be run entirely in the browser or self‑hosted via Docker. Supports many AI and search providers, offers client‑side (keys in the browser) and server‑side (keys in env vars) modes, and exports results as Markdown or PDF.

  • Project
    DashAISoftware/dashAI

    dashAI is a cross‑platform graphical toolbox for training, evaluating and deploying modern AI models (especially LLMs). It offers ready‑to‑run installers, a PyPI package, and Docker images, supports CPU‑only as well as CUDA/ROCm GPU acceleration, and includes experiment management, dataset handling, and database migrations.

  • Project
    NSHipster/sosumi.ai

    sosumi.ai is a Node‑based service that converts Apple Developer documentation (including Swift docs, HIG, and WWDC transcripts) into clean Markdown, making the content easy for AI models and automation tools to consume. It offers a simple URL rewrite, an HTTP/MCP API, a CLI, a Chrome extension, and can be self‑hosted on Cloudflare Workers or any Hono‑compatible runtime.