Federated queries, one window.
HTTP statement connections with catalog, schema, and table browsing, forward-only paging and cancellation, EXPLAIN, and import/export — across every catalog your coordinator can reach.
Trino & Presto · HTTP · catalogs
The plan Mongrel would build for your query.
Click any stage below to see rows, CPU, and what Mongrel would suggest. Same distributed-plan view you get inside the workbench.
// ScanFilter · iceberg.analytics.events
Scans all 412 ds partitions and applies event_type = 'purchase' as a residual filter. 2.1B rows read, 388s of CPU. This is the bottleneck.
-- trino-coordinator · catalogs: iceberg, mysql SELECT a.account_name, SUM(e.amount) AS total FROM iceberg.analytics.events e JOIN mysql.crm.accounts a ON a.account_id = e.account_id WHERE e.event_type = 'purchase' GROUP BY a.account_name ORDER BY total DESC LIMIT 20; -- 6m 42s · 412 ds partitions, 2.1B rows scanned -- ✓ Proposal: prune the ds partition before the join WHERE e.ds >= DATE '2026-07-01' AND e.event_type = 'purchase'
What Mongrel ships for Trino.
Catalog browser
Drill catalog, schema, and table across every connector your coordinator exposes — Hive, Iceberg, Delta Lake, MySQL, PostgreSQL.
Federated SQL editor
One editor for cross-catalog queries. Dialect-aware completion, catalog.table resolution, and join hints across connectors.
EXPLAIN & stage stats
Run EXPLAIN logical or distributed. Walk fragments, exchanges, and operators with row counts and CPU per stage.
Forward-only paging
Stream millions of rows through nextUri paging without loading the result set. Memory stays flat on the longest scans.
Query cancellation
Kill any running statement from the grid — Mongrel issues the DELETE to the coordinator and the workers unwind cleanly.
Trino & Presto profiles
The same HTTP statement protocol, tuned per engine. Connection profiles for Trino and Presto coordinators, TLS and basic auth.
Table stats & partitions
SHOW STATS, partition listings, and column metadata at a glance. See what a predicate will actually prune before you run it.
AI explain & rewrite
Paste a query and an EXPLAIN. The assistant proposes partition predicates, join reordering, and connector pushdowns.
Import, export, copy
Export result sets to CSV, JSON, or Parquet. Copy data between catalogs — or between Trino and any other engine.