TableStudio is a native Mac client for Cloudflare D1. Cloudflare's own tools are wrangler and the dashboard. TableStudio gives you a real table browser and SQL editor for your D1 databases, backups on your Mac, and a safe way in for coding agents like Claude Code, Codex and Cursor.
Back up D1 to your Mac
D1 has no dump tool you can run locally. TableStudio copies a database's schema and rows over the D1 API into a SQLite file, on your Mac or in any S3 bucket. Before a risky change, take a restore point and roll back to it if you need to.
Want to try something without touching the real database? A sandbox copies it into a SQLite file on your Mac, with its real rows or none, so you can experiment, or let an agent experiment, offline.
What TableStudio does for D1
A desktop home for the databases behind your Workers.
Agents in your Workers projects
- Claude Code, Codex, Cursor and OpenCode reach D1 by name, with no API token in a config file.
- Agent reads are checked by their SQL before they run.
- Production changes are staged for your review and applied with Touch ID.
- The rows an agent is about to touch are copied first, so one bad change can be put back.
- Values the agent reads from D1 are marked as untrusted, so a row can't pass itself off as instructions.
Sign in, not set up
- Sign in with Cloudflare and import every D1 database in the account.
- Or use an API token with D1 edit access, kept in the macOS Keychain.
- Find the D1 databases named in your projects'
wranglerfiles.
Query Insights from D1 analytics
- The statements your database spends its time on, read from Cloudflare's own D1 analytics.
- What's wrong with each, in plain words.
- Widget packs for D1's health and schema objects on the dashboard.
Schema, SQLite-style
- Change a column, and TableStudio rebuilds the table the way SQLite requires, after a restore point.
- Views, triggers, indexes and foreign keys.
- Schema checks for loose SQLite typing, duplicate indexes and foreign keys without one.
- Watch your Drizzle or Prisma migrations folder: pending, applied and drifted.
Browse and edit
- Filter, sort and search, and hover a foreign key to preview its row.
- Edits apply one by one, since D1 has no transactions to group them.
- CSV and TSV import; export to CSV, TSV, JSON, SQL, Markdown or Excel.
- Fake rows that follow your foreign keys, for a staging database that looks real.
Move and compare
- Compare two D1 databases, say production and staging, and get the script that closes the gap.
- Move a D1 database into Postgres or MySQL with a pre-flight report.
- Sync rows between databases, copying only what changed.
What D1 can't do
D1 runs SQLite behind an HTTP API, which shapes what any client can offer. There are no transactions, so a batch of edits applies one by one and a dry run can name a statement but not count its rows. There's no interactive shell, so there's no "Open in Terminal". TableStudio is upfront about each of these in the app, rather than pretending D1 is a local SQLite file.
Last updated