SnowSQL is Snowflake's official command-line tool — powerful but text-only. Snowsight is the official web UI — comprehensive but slow on large queries. QueryFlow is the native Mac desktop client that combines fast interactive querying with Claude AI assistance and scheduled jobs.
Quick answer: QueryFlow is the native macOS alternative to SnowSQL (Snowflake's CLI) and Snowsight (Snowflake's web UI). It provides a desktop SQL editor with Claude AI, multi-tab workspaces, scheduled queries, Python notebooks via Flow Books, and Visual ETL pipelines. Uses Snowflake's SQL API v2 with Programmatic Access Token authentication. $299.99/year.
Snowflake users have four main options on Mac. Snowsight: the official web UI, comprehensive but slow on large result sets, requires browser. SnowSQL: the official CLI, scriptable but text-only, no syntax highlighting or autocomplete by default. DBeaver: a Java-based universal SQL tool with Snowflake support via JDBC, runs on JVM with all the associated cold-start and memory issues. DataGrip: JetBrains' commercial SQL IDE, also Java-based, $118/year. QueryFlow is the only purpose-built native Swift Snowflake client.
Snowflake operations are interactive — you write a query, see results, refine the query, repeat. The faster the loop, the more productive the work. Browser-based Snowsight has noticeable latency on result rendering and tab switching. JVM-based clients have 5-15 second cold starts. SnowSQL is fast but you give up the visual feedback. QueryFlow's pure Swift implementation launches in under a second and renders results in a virtualized table that handles hundreds of thousands of rows without lag.
Snowflake's recommended authentication for desktop clients in 2026 is Programmatic Access Tokens (PATs). QueryFlow uses PATs natively — generate a token in Snowsight, paste it into QueryFlow once, and the connection is set up. The token is stored encrypted in macOS Keychain. No JDBC drivers, no Java runtime, no key file management like SnowSQL requires for key-pair auth, no SSO browser dance that slows down quick connection switches.
Once connected to Snowflake, QueryFlow's Claude AI integration knows every database, every schema, every table, every column. Ask Claude to write a query against your FACT_ORDERS table joined to DIM_CUSTOMERS and it writes valid Snowflake SQL with your actual column names. Ask why a query is slow and Claude analyzes the query profile (Snowflake's EXPLAIN equivalent). Ask to convert a subquery to a CTE and it rewrites in Snowflake-idiomatic style with WITH clauses.
QueryFlow supports unlimited query tabs. Each tab maintains its own SQL editor state with full Snowflake dialect syntax highlighting (FLATTEN, LATERAL, PIVOT, QUALIFY, MATCH_RECOGNIZE, time travel, zero-copy clone). Tabs are auto-named from your FROM clause and restored on next launch. SnowSQL has no concept of persistent tabs.
Any Snowflake query becomes a scheduled job with one click. Pick a trigger (daily 6 AM, hourly, custom cron). Pick a destination (write back to Snowflake as a new table, email the result, upload to S3, push to Google Sheets, send to Salesforce). The scheduler runs locally on your Mac via SMAppService. For workflows that don't need Snowflake Tasks' compute (which costs Snowflake credits), this is dramatically cheaper.
Beyond ad-hoc querying, QueryFlow's Visual ETL builder treats Snowflake as a first-class source or destination. Build a Snowflake-to-Salesforce reverse ETL pipeline in 5 minutes. Sync a Postgres production database to Snowflake. Pull Salesforce data into Snowflake daily. All without leaving the same app you write SQL in.
Yes. QueryFlow's SQL editor supports parameter substitution at query time. You can define parameters with default values and override them per execution.
QueryFlow covers all the common interactive query work plus scheduled execution. SnowSQL's specific batch-script execution model (running a .sql file with !variable directives) is more naturally handled in QueryFlow as a Flow Book with SQL cells.
PUT and GET stage operations are not currently first-class commands in QueryFlow's editor (they're SnowSQL-specific commands, not standard SQL). For file staging operations, SnowSQL remains the right tool. Most users use stages indirectly through COPY INTO statements, which QueryFlow handles fine.
No. QueryFlow only consumes Snowflake credits when you actively run queries. Idle warehouse time is governed by your warehouse's AUTO_SUSPEND setting (default 5 minutes). QueryFlow doesn't run background polling or keepalive queries.
QueryFlow v1.5 uses Programmatic Access Tokens (PAT) as the primary authentication. Key-pair authentication is on the roadmap. PAT is Snowflake's recommended modern authentication and works for all current desktop client use cases.
14-day free trial. Connect Snowflake with a PAT in 60 seconds and feel the difference from Snowsight or DBeaver.