The TiDB workbench, end to end.
MySQL-protocol SQL with schema browsing, row and bulk CRUD, EXPLAIN plans across TiDB and TiKV, transactions and savepoints, import/export, and the shared MySQL-family workbench tools.
MySQL protocol · TiDB · TiKV
The plan Mongrel would build for your query.
Click any node below to see row estimates, which layer executes it, and what Mongrel would suggest. Same view you get inside the workbench.
// IndexLookUp · orders, cop[tikv]
Double read on TiKV: index range scan collects handles, then a second coprocessor task fetches full rows by handle. The expensive part of this plan.
-- orders · tidb-prod:4000 SELECT o.id, o.total, o.status, u.email AS customer_email FROM orders o JOIN users u ON u.id = o.user_id WHERE o.created_at > '2026-06-01' AND o.status = 'paid' ORDER BY o.created_at DESC LIMIT 20; -- 412ms · IndexLookUp double read on TiKV -- ✓ Proposal: composite (user_id, created_at) ALTER TABLE orders ADD INDEX idx_user_created (user_id, created_at);
What Mongrel ships for TiDB.
Schema browsing
Browse databases, tables, indexes, and views over the MySQL protocol. Generated columns and clustered indexes included.
Row & bulk CRUD
Edit rows inline, stage multi-row inserts, updates, and deletes, and preview the statements before they run.
EXPLAIN & EXPLAIN ANALYZE
Inspect root vs cop[tikv] tasks, estRows vs actRows, and where TiKV does the work. Plans render as clickable trees.
Transactions & savepoints
BEGIN, SAVEPOINT, ROLLBACK TO, COMMIT — with a clear view of optimistic vs pessimistic transaction mode per session.
TiKV plan inspection
See TableReader, IndexReader, and IndexLookUp operators, coprocessor pushdown, and when a double read is costing you.
Index proposals
Mongrel proposes composite indexes from your slow queries and shows the predicted plan before you apply anything.
AI explain & rewrite
Paste a query and its EXPLAIN. The assistant rewrites it, proposes indexes, and explains the TiDB planner's choices.
Users, roles & TLS
Manage MySQL-compatible users, roles, and grants. Connect with TLS client certs to TiDB Serverless or self-hosted clusters.
Import, export, copy
CSV, JSON, and SQL dumps in and out. Copy data between TiDB and any other engine — schema-aware.