Supabase's web SQL editor is great for quick queries. For real day-to-day Postgres work — long queries, multi-tab debugging, scheduled jobs, AI assistance — QueryFlow is the native Mac client.
Quick answer: QueryFlow is a native Mac client for Supabase Postgres databases. It connects via standard Postgres wire protocol with SSL using your Supabase project's connection string. Features include Claude AI with full schema awareness, multi-tab workspace, scheduled query execution, and Visual ETL pipelines. $299.99/year.
Supabase's built-in SQL editor in the dashboard is genuinely good for quick queries — autocomplete, syntax highlighting, save history. For occasional use it's perfect. The friction starts when you do real day-to-day work: queries longer than a few lines feel cramped, switching between multiple in-progress queries means losing state, debugging complex joins requires more screen real estate, and there's no scheduling for recurring work.
From your Supabase project dashboard: Settings → Database → Connection string. Copy the URI mode connection string (it'll look like postgresql://postgres:[YOUR-PASSWORD]@db.PROJECT.supabase.co:5432/postgres). In QueryFlow: add a new Postgres connection and paste the URI. SSL is enabled automatically. Connection is set up in 30 seconds. The database password is the one you set when creating the project (different from your Supabase dashboard password).
Supabase provides both direct connections (port 5432, for long-lived sessions) and a Supavisor-pooled connection (port 6543, for many short-lived connections). For QueryFlow's desktop-client use case, the direct connection is the right choice — you have a single user with a single connection, and you benefit from full Postgres features (prepared statements, session-level settings, etc.).
Open separate tabs for separate concerns: one for the query you're debugging, one for the schema you're exploring, one for the migration you're prototyping, one for the analytics question you're investigating. Each tab maintains its own SQL editor state, syntax highlighting, query history, and result set. Close QueryFlow, reopen later, every tab is right where you left it.
Claude AI is integrated into the SQL editor with full Supabase Postgres schema awareness. After connection, Claude sees your tables (including Supabase's auth.users, profiles, RLS-enabled tables), columns, types, and relationships. Ask Claude to write a query joining auth.users to your profiles table and it produces valid SQL with the correct relationship. Ask why a query is slow and Claude analyzes EXPLAIN output.
Supabase relies heavily on Postgres Row-Level Security (RLS) for client-side security. When QueryFlow connects with the postgres role (the database owner), RLS is bypassed (the role has BYPASSRLS). For testing your application's actual RLS behavior, you'd connect with a non-superuser role. Most QueryFlow users connect as postgres for development/analytics work and use Supabase's client libraries (with anon/authenticated roles) for application code.
Supabase Postgres can be a source for ETL pipelines. Common patterns: sync Supabase user data to Salesforce for sales operations, export daily activity to Google Sheets for the product team, replicate to Snowflake for analytics. QueryFlow handles all of this with scheduled pipelines running locally on your Mac.
Yes. The auth schema and its tables (auth.users, auth.identities, etc.) are visible in QueryFlow's schema explorer just like any Postgres schema. You can query them, join them with your application tables, and include them in pipelines.
QueryFlow's SQL editor can execute any DDL (CREATE TABLE, ALTER TABLE, etc.) so manual schema migrations work fine. For migration-tool workflows with version tracking and rollback, the Supabase CLI or tools like Drizzle Kit remain the right choice.
No. QueryFlow connects at the Postgres database level via wire protocol. Supabase's REST (PostgREST) and GraphQL APIs are separate layers that QueryFlow doesn't interact with.
Storage objects (files) live outside Postgres. QueryFlow can query the storage.objects table to see metadata about stored files, but actual file content is in S3-compatible storage that QueryFlow doesn't currently interact with directly.
Yes. Each Supabase project gets its own QueryFlow connection. The connection list shows all of them, and switching between them is a click.
14-day free trial. Connect to your project in 30 seconds and gain Claude AI plus the desktop SQL workflow.