IDE Tool Overview
The IDE is a browser-delivered, server-side database workbench: connection state, pooling, introspection, and query execution all live on the backend, and the client is a thin Monaco-based front end talking to it over the Next.js → Express proxy. Nothing installs on the operator's machine, and no database driver, credential, or network path ever touches the browser, the server holds the pools and speaks each engine's native protocol on your behalf. One workbench spans relational, document, key-value, wide-column, graph, search, time-series, and analytical engines with a consistent editor, navigator, and results surface, while surfacing each engine's real capabilities rather than flattening them to a lowest common denominator.
Workbench Surfaces
Connection Manager
Per-user, server-persisted connection registry. Connections are scoped to your account (not shared across users), survive sign-out and browser-cache clears, and can be arranged into a client-side folder hierarchy for environment/project/team segmentation. Each connection carries its own pool and can back one or more editor tabs concurrently. Starter licenses cap the registry at 10 connections; Professional and Enterprise are unlimited. Connections can be imported from the DB DevOps Manager as independent copies (see Connecting to Databases).
Object Explorer / Navigator
Lazy-loaded, engine-aware introspection tree: namespaces, object groups, columns with type/nullability/default/key role, indexes, constraints, and, where the engine exposes them, server-level nodes (server info, sessions, locks, users, roles, tablespaces). The panel is always docked on the right — collapsed to a slim rail until you need it, expanding automatically on your first connect. The navigator is a stacked pair, object tree above, an inspector/definition pane below, with a draggable, persisted divider. Covered in depth under Schema Explorer.
Query Editor
Monaco with per-dialect language services: dialect-correct keyword/function highlighting, schema-aware completion, formatting, and multi-statement execution. Tabs are independent SQL documents, each bound to a connection, and both open tabs and unsaved buffer text are persisted per user across sessions and across an IDE↔DevOps context switch.
Results & Analysis
A virtualized/paged grid for row-returning statements, affected-row and server-message reporting for DML/DDL, an execution-plan surface (EXPLAIN / EXPLAIN ANALYZE) on engines that support it, inline charting, and export to CSV/JSON/SQL INSERT/clipboard. A rolled-back-transaction preview reports the blast radius of a mutating statement without committing it.
AI Assistant
A conversational assistant (requires an AI-enabled license and a configured provider) grounded in the connected database's live schema. It holds a real multi-turn conversation: ask a follow-up like "why did that return 23?" and it reasons about the query and results from earlier in the thread. Generated queries arrive in a code block you can insert into the editor or run directly; the destructive-statement guard still applies. Each user's conversation is retained locally and restored across page refreshes, sign-out/in, and switching between the IDE and DB DevOps, clear it any time from the panel header.
Switching between the IDE and DB DevOps preserves your working state on both sides, selected connection, open results, and in-progress forms are exactly where you left them when you switch back.
Supported Engines
The connection type picker is a searchable typeahead grouped by engine family. Introspection depth tracks what each family's catalog actually exposes:
| Family | Engines | Introspection depth |
|---|---|---|
| Relational | PostgreSQL, MySQL, MariaDB, SQL Server, Oracle, SQLite, CockroachDB | Full catalog: namespaces, tables/views/materialized views, functions/procedures, sequences, indexes, constraints, triggers; column type/nullability/default/key role; object DDL scripting; server info, sessions, locks, users/roles, tablespaces where the engine exposes them. |
| Time-series | TimescaleDB | Full relational introspection plus hypertable awareness layered over the PostgreSQL catalog. |
| Analytical | ClickHouse (+ cloud warehouses surfaced in the picker: Snowflake, Redshift) | Databases/schemas, tables and columns, engine-specific object metadata. |
| Document | MongoDB | Databases → collections; field structure inferred from sampled documents (schemaless, so shape is derived, not declared). |
| Key-value | Redis | Keyspace browsing by pattern, per-key type and TTL, logical DB index selection. |
| Wide-column | Cassandra | Keyspaces, tables, and CQL-level column metadata. |
| Graph | Neo4j | Labels, relationship types, and property keys. |
| Search | Elasticsearch | Indices and mappings. |
Selecting a family member pre-populates engine-appropriate connection defaults; anything you've already overridden is preserved. Capability differences (completion, EXPLAIN, DDL scripting, charting/export) are surfaced per dialect rather than assumed uniform, see the capability matrix under Keyboard Shortcuts & Multi-Database Support.
Layout
- Left rail, connection registry over the schema navigator. Toggle with
Ctrl/Cmd+B. - Center, tabbed editor; each tab is an independent document with its own connection context.
- Bottom, results grid, messages, execution plan, and chart tabs.
- Right navigator, stacked object tree and inspector/definition panes with a draggable divider (double-click to reset). All panel sizes and the divider position persist per session.