Neon is the serverless Postgres many modern stacks have adopted. QueryFlow connects to any Neon project, including branches, via standard Postgres wire protocol with SSL — and adds Claude AI in the editor.
Quick answer: QueryFlow is a native Mac client for Neon Postgres. It connects to any Neon project — including individual branches — via standard Postgres wire protocol with SSL. Features include Claude AI with schema awareness, multi-tab workspace, scheduled queries, and Visual ETL pipelines. Built in pure Swift for $299.99/year.
Neon is the serverless Postgres offering most modern stacks are choosing. Key features: branching (instant database copies for development), scale-to-zero pricing, and standard Postgres compatibility. The branching workflow is especially powerful — every git branch can have its own database branch, enabling true preview environments. QueryFlow handles all of this cleanly.
From the Neon console: copy your connection string from the Dashboard. In QueryFlow: add a new Postgres connection, paste the connection string. QueryFlow parses host, port, user, password, database, and enables SSL (Neon requires SSL). For branch connections, the connection string changes per branch — add each branch as its own QueryFlow connection with a clear name (myapp-main, myapp-feature-x).
Neon's branching makes it practical to have separate database environments per development branch. Pattern: create a Neon branch for your feature, add the new connection string to QueryFlow, develop with isolated data, run migrations against the branch, validate, then merge. QueryFlow's connection list shows all branches at a glance — switch between them in one click.
Neon's serverless model means the first query after idle has a cold-start penalty (usually 0.5-2 seconds). QueryFlow handles this gracefully — the loading state is clear, the connection retries are transparent. For interactive query work, the cold start is rarely noticeable because you're typing the next query during the warmup. For scheduled jobs, the cold start is part of the run time but doesn't cause failures.
Claude AI in QueryFlow's editor sees every table, column, type, and relationship in your Neon project. Ask Claude to write a query against your users table joined to subscriptions, and it produces valid Postgres SQL with your actual column names. Ask about query optimization and Claude reads the EXPLAIN ANALYZE output to suggest index additions or query restructuring.
QueryFlow's Visual ETL treats Neon as a first-class source or destination. Common patterns: sync Neon production data to Snowflake for analytics, export specific table snapshots to S3 for backup, pipe data from Neon to Salesforce for sales operations. Schedule these to run daily, no orchestration infrastructure required.
Neon connection strings include credentials. QueryFlow stores them in macOS Keychain with Secure Enclave protection on Apple Silicon Macs. Connection strings are never transmitted to QueryFlow's servers (QueryFlow has no servers handling your queries — everything is direct client-to-database). For organizations with security requirements, this client-side architecture is meaningfully better than cloud-based SaaS database clients.
Each Neon read replica has its own connection string. Add each as a separate QueryFlow connection. For read-heavy analytical queries on production data, pointing QueryFlow at a read replica reduces load on the primary endpoint.
QueryFlow is a client — it queries whatever compute Neon provisions. Autoscaling happens on Neon's side based on your project configuration. QueryFlow's query execution behavior doesn't change.
Yes. Neon provides both direct and pooled connection strings. Pooled connections work fine with QueryFlow for typical query work. Note that pooled connections have some Postgres feature limitations (no prepared statements across queries, transaction-pooled mode differences) that may matter for specific advanced workflows.
Not as first-class UI in v1.5. The connection settings page shows the standard Postgres connection info. Neon-specific metadata (compute units, branch parent, last activity) remains in the Neon console.
Schema migrations can be executed as SQL in QueryFlow's editor or as a Flow Book with multiple migration steps. For migration-tool-style workflows (with rollback, version tracking, etc.), tools like Flyway, Alembic, or Prisma remain the right choice.
14-day free trial. Connect to any Neon project or branch in under a minute and get Claude AI in your SQL editor.