Free PostgreSQL Client for Windows,
Mac (Apple Silicon) & Linux

Visual EXPLAIN ANALYZE

Estimated plans run EXPLAIN (FORMAT JSON) on any PostgreSQL version; Analyze runs EXPLAIN (ANALYZE, BUFFERS, FORMAT JSON), so buffer hit/read counts are always captured, not an opt-in extra. On a PostgreSQL 19 (beta) server the capture adds the IO option automatically for per-node I/O timing. Cost and percent-of-total badges are read straight from each node's own Total Cost value, and you can compare two plans side-by-side.

Schema & Data Compare

Diff schemas and table data between two PostgreSQL databases. View side-by-side DDL differences and generate synchronization scripts to keep environments in sync.

AI-Native & Local-First

Jam SQL Studio combines an AI Workspace (file-backed tabs + context) with a local MCP server (Model Context Protocol), so tools like Claude Desktop / Claude Code, OpenCode, and Codex CLI can safely call SQL Studio as a database tool under your AI policy.

DBA Tools

The Session Browser reads pg_stat_activity and pg_locks for live sessions, wait events, and lock contention, and can terminate a session via pg_cancel_backend / pg_terminate_backend. The Performance Dashboard adds top-SQL-by-time, wait statistics, and (PostgreSQL 19, beta) an autovacuum scoring panel — all without leaving the IDE.

PostgreSQL's JSONB and Enum Columns, Done Right

PostgreSQL has some of the deepest JSON support of any relational engine, so Jam SQL Studio treats jsonb columns as first-class instead of raw text. The JSON filter chip covers key presence, equality/inequality, numeric comparisons, contains, empty checks, and array-wildcard ([*]) matching, built on #>> / #>, jsonb_path_query, and jsonb_path_exists. Recursive-descent paths ($..foo) are supported natively (rewritten to jsonpath's **) — PostgreSQL and Oracle are the only two engines where that works. On PostgreSQL 19 (beta) the chip also gains case-insensitive equality (path_ieq / any_ieq) via the jsonpath .lower() method, gated to PG19+ because older servers don't have it.

Declare a column as an enum and Jam SQL Studio sources its values in priority order from the native pg_enum type, a column CHECK (col IN (...)) constraint, or a sampled SELECT DISTINCT — capped at 200 distinct values (scanning at most 100,000 rows, 5-second wall-clock cap) before the picker falls back to free text. See the Enum Columns and JSON Columns guides for the full mechanics.

PostgreSQL 19 (Beta): What's New in Jam SQL Studio

The Object Explorer's table context menu gains a Maintenance submenu (REPACK, REPACK CONCURRENTLY, VACUUM, ANALYZE — REPACK is 19's successor to VACUUM FULL/CLUSTER). Security Manager's Script as CREATE for roles and databases uses the native pg_get_role_ddl / pg_get_database_ddl functions when present. SQL/PGQ property graphs are listed in the Object Explorer with GRAPH_TABLE query scaffolding. Six built-in SQL snippets cover REPACK, WAIT FOR LSN, ON CONFLICT DO SELECT, temporal FOR PORTION OF UPDATE/DELETE, and GROUP BY ALL. Everything here falls back gracefully on older servers — nothing requires PG19 to use the app.