QueryFlow is a pure Swift desktop app built around the Snowflake SQL API v2. PAT authentication, schema-aware Claude AI, Python notebooks, and scheduled jobs that run locally on your Mac.
Snowflake's web interface (Snowsight) is solid, but it is still a browser tab. For people who spend most of their day inside Snowflake, a native desktop app is faster, integrates with system shortcuts, and never loses your work to a browser crash. QueryFlow is the first Mac App Store application built around Snowflake as a primary use case.
QueryFlow connects to Snowflake using the modern SQL API v2 with Programmatic Access Token (PAT) authentication — the same security model Snowflake recommends for desktop applications. You generate a PAT in your Snowflake settings, paste it into QueryFlow once, and the token is stored in the macOS Keychain. The connection auto-resumes your warehouse, respects your role, and handles refresh transparently.
Because QueryFlow is pure Swift and SwiftUI, the experience feels like a first-party Apple app. Trackpad gestures work natively. The menubar is integrated. The clipboard handles results correctly. Tabs are restored between sessions. Liquid Glass surfaces match macOS Tahoe. Dark mode is the default and looks intentional, not retrofitted from a web theme.
When you connect QueryFlow to Snowflake, Claude AI automatically sees your databases, schemas, tables, and columns. Ask Claude to write a query against your TPCH_SF1 customer table and it writes valid SQL with the correct column names — C_CUSTKEY, C_NAME, C_MKTSEGMENT, C_NATIONKEY. Ask Claude why a JOIN is slow and it reads the execution plan and explains the bottleneck. The AI never sees your data unless you explicitly send query results — only schema metadata flows by default.
Every Snowflake account ships with SNOWFLAKE_SAMPLE_DATA — TPCH and TPCDS test datasets at multiple scale factors. QueryFlow detects these and surfaces them in the schema explorer immediately. If you are evaluating Snowflake performance, evaluating QueryFlow, or learning SQL, these datasets are the easiest way to see real query work in a real-sized warehouse without spending a dollar on compute beyond Snowflake's free trial credits.
Once your Snowflake query works, QueryFlow lets you schedule it. Cron, interval, daily, weekly, or custom — the scheduler runs locally on your Mac. The job authenticates to Snowflake, runs the query, and delivers results to Email, S3, Google Sheets, SFTP, a local file, or back into Snowflake itself as a new table. Zero cloud orchestration infrastructure. Zero AWS Glue. Zero Lambda. The only compute cost is whatever Snowflake charges for the warehouse usage during the query.
QueryFlow's Data Sync feature lets you map fields visually between any source and any destination. Bring data into Snowflake from PostgreSQL, MySQL, Salesforce, Google Sheets, or CSV files. Hit AI Map and QueryFlow matches source columns to Snowflake table columns using 25 synonym groups (it knows email_addr maps to EMAIL, fname maps to FIRST_NAME, phone_cell maps to MOBILE_PHONE). The mapping runs entirely on-device — no field names leave your Mac during the matching step.
Snowflake's official Snowsight interface runs in a web browser. For a native macOS experience, QueryFlow is the most modern option — written in Swift and SwiftUI, connecting to Snowflake via the SQL API v2 with Programmatic Access Token (PAT) authentication. The connection auto-resumes the warehouse and respects all Snowflake roles and policies.
QueryFlow uses Snowflake's Programmatic Access Token (PAT) authentication, which is the modern recommended approach for desktop applications. You generate a PAT in your Snowflake account settings, paste it into QueryFlow once, and it is stored encrypted in the macOS Keychain. Snowflake allows up to 15 PATs per user, so you can have separate tokens for different machines or environments.
Yes. Any Snowflake account includes the SNOWFLAKE_SAMPLE_DATA shared database with TPCH and TPCDS test datasets at multiple scale factors. QueryFlow can browse and query these out of the box once you connect — no additional setup required. It is a good way to test the editor and Claude AI before pointing QueryFlow at your production data.
Yes. Each connection can specify a role and a warehouse. QueryFlow respects the policies of whichever role you authenticate as. You can switch warehouses per connection if you maintain separate compute resources for different query types — for example a small warehouse for exploration and a larger one for scheduled ETL jobs.
Yes. QueryFlow's Data Sync and Scheduler features can write query results back to Snowflake tables as Insert, Update, or Upsert operations. You can also use QueryFlow's AI Map feature to map columns from one Snowflake table to another, or from a different source like Postgres into Snowflake automatically.
Stop living in a browser tab. QueryFlow is on the Mac App Store with a 14-day free trial. Connect your Snowflake account and run your first query in under 60 seconds.