DBcooper

A database client for PostgreSQL, MySQL, MariaDB, SQLite, DuckDB, Redis, ClickHouse, and Cloudflare D1, built with Tauri, React, and TypeScript.

brew install --cask --force amalshaji/taps/dbcooper

Homebrew clears the Gatekeeper quarantine automatically, so the app opens right away.

Download the latest .dmg from Releases.

macOS users: After installing (before opening the app the first time), bypass Gatekeeper since the app isn't notarized:

xattr -cr /Applications/DBcooper.app

Then you can open the app normally.

Check out the full list of features on our documentation site.

From the Connections screen, DBcooper can:

• Create PostgreSQL 17, MySQL 8.4, MariaDB 11.4, Redis 7, or ClickHouse 25.8 in a Docker container and save a ready-to-use connection.
• Link a compatible PostgreSQL, MySQL, MariaDB, Redis, or ClickHouse container that already exists in the current Docker context, including Docker Compose services.
• Copy the complete connection string for a Docker-managed connection.
• Stop and restart the linked container from the connection menu.

Databases created by DBcooper use a persistent named volume. Quitting DBcooper stops the container without removing the container or its data, and opening the connection starts it again. Deleting a connection also preserves Docker data unless you explicitly choose to remove it.

Linking an existing container requires its database port to be published to the host. DBcooper reads standard image environment variables, including Docker secret-style *_FILE values, and lets you review the detected connection fields before saving.

Find answers to common questions on our documentation site.

• Frontend: React + TypeScript + Vite
• Backend: Rust + Tauri v2
• Database: SQLite app storage; PostgreSQL, MySQL, MariaDB, SQLite, DuckDB, Redis, ClickHouse, and Cloudflare D1 connections
• UI: shadcn/ui components
• Package Manager: Bun

• Bun - JavaScript runtime and package manager
• Rust - For Tauri backend
• macOS 26 (Tahoe) or later

# Install dependencies
bun install

# Run in development mode
bun run tauri dev

# After changing icon.png or src-tauri/macos/AppIcon.png
# regenerate every platform icon and the macOS asset catalog
bun run icons

# Build for production
bun run tauri build

# Test D1 against a disposable local Wrangler database
bun run test:d1-local

DBcooper includes an opt-in, token-authenticated MCP server for external coding agents. See docs/mcp.md for setup and curl examples.

To create a D1 connection:

1. Open the Cloudflare dashboard Account home, find the account, open its menu, and choose Copy account ID. See Cloudflare's Account ID guide.
2. Go to My Profile → API Tokens → Create Token → Create Custom Token. Under permissions, choose Account → D1 → Read for browsing and querying, or Edit if DBcooper should edit rows or create tables. Scope the token to the account from step 1. See Cloudflare's API token guide.
3. Create the token and copy its secret immediately. Cloudflare shows the secret only once, so store it securely.
4. In New Connection, choose Cloudflare D1, enter the Account ID and token, select Load databases, then choose a database from the list.

The in-app connection form links to the same Cloudflare D1 setup guide.

D1 connections use Cloudflare's HTTPS API directly and do not support SSH tunnels. Cloudflare applies a global API limit of 1,200 requests per five minutes per user; DBcooper surfaces rate-limit responses and does not automatically replay D1 requests. Because the API transports results as JSON, integers outside JavaScript's safe integer range may lose precision when displayed.

To use AI-powered SQL generation:

1. Go to Settings (gear icon) and choose an AI provider:

   • OpenAI-compatible API: configure an API key, endpoint, and model.
   • Claude Code, Codex CLI, or opencode: DBcooper uses your local CLI install and existing login.

2. In the Query Editor, you'll see an instruction input above the SQL editor:

   • Type a natural language description (e.g., "show all users with posts from last week")
   • Click Generate or press Enter
   • Review the generated SQL before running it

The AI receives your instruction, existing editor SQL, and selected table/column schema metadata for accurate query generation.

The app is configured to build for macOS ARM (Apple Silicon). The build process:

1. Creates optimized production bundles
2. Signs the app with your signing key
3. Generates updater artifacts

Releases are automated via GitHub Actions. To publish a new version:

1. Update version in src-tauri/tauri.conf.json
2. Open a PR and add the release label
3. Merge the PR into main
4. GitHub Actions will create and push the tag (e.g., v0.0.42), then build a draft release
5. Review and publish the release

Every commit merged into main publishes a signed canary after the test suite passes. Canary versions use the next patch version with a build suffix, such as v0.0.64-canary.142.

Canary updates are disabled by default. To receive them, open Settings and enable Canary updates. Disable the setting to return to the stable channel; if the installed canary is newer than the latest stable release, DBcooper will wait for the next newer stable release instead of downgrading.

Set these in your GitHub repository settings:

• TAURI_SIGNING_PRIVATE_KEY - Contents of your signing key file
• TAURI_SIGNING_PRIVATE_KEY_PASSWORD - Password (if set)

MIT