What's new

[1.2.1] — 2026-10-06

Fixed

• AI with Gemini: masking rules, "Which fields are PII?", "Build dashboard" and "Explain indexes" now work (Gemini rejected their requests before). AI-built pipelines no longer fail when Gemini escapes $ in a stage.
• AI backup models now take over when the main model hangs, not only when it returns an error, and Test connection names the model that actually answered.
• Query Builder: the sort reaches Explain and Query Code, Add Condition adds one row, and tab titles count per kind.
• Aggregation: field pickers list the collection's fields, including arrays for $unwind.
• Document history: dates show correctly, revert works, and the selected entry follows the theme.
• Compare & Sync: the sync preview is visible again.
• Collection lists refresh after a copy; the grid refreshes after a bulk edit and counts update after an insert.
• Data masking: number noise keeps the original decimals, and "Already exists" no longer shows after a write.
• Test data: values seen in arrays become choices for generated data.
• Create Index form layout, Index Review keeps its results across tab switches.
• Server Monitoring: memory is shown and the trend charts fit their cards.
• Dashboards: chart colours stay the same per category, and the chart editor and toolbar fit.
• Tasks: the editor background, the keychain notice and the Counts column.
• SQL Migration can use saved PostgreSQL and MySQL connections, including SSH tunnels.
• No more stray tooltips in the SQL editor and results grid.

[1.2.0] — 2026-10-05

Added

• Ask AI on nearly every screen: the chart editor and dashboard builder, profiler, server monitoring, index review, import, export, test data, bulk edit, value search, the document editor, roles, the task scheduler, the data model, schema compare, connection errors, the SQL editor, and "Ask AI about this screen" in the command palette. Each helper suggests something you then apply; nothing runs on its own, and every helper follows the connection's "Allow AI" policy.
• Gemini backup models: when Gemini is busy or rate-limited, MotionQL retries on a backup model and says which model answered (Settings → AI → Backup models).

Changed

• Help text and secondary panels stay hidden until you ask for them.
• Dialogs close with Escape and keep keyboard focus inside them.

Fixed

• The sidebar's Export and Import menu items now open the right tools.

[1.1.0] — 2026-10-04

Added

• SQL client for PostgreSQL, MySQL, MariaDB, SQL Server and Oracle: browse objects, run SQL, edit table rows.
• Split panes and tab groups, with live results in every pane.
• Real Server Monitoring (serverStatus, dbStats, top, server log) and a real explain with index suggestions in the visual Query Builder.
• Import connections from Compass, Studio 3T and Robo 3T.
• Edit users and change passwords; GridFS previews and drag-and-drop upload.
• Dashboards shared through the workspace folder or the Team Server.
• A guide to getting a free Gemini API key.
• Linux AppImage and deb packages.

Changed

• Every screen was redesigned to be more compact. Secondary panels (previews, stats, help, history, rarely used options) open and close with a button, and each screen remembers your choice. The sidebar can be hidden with Ctrl/Cmd+Alt+B.
• Faster results: the first rows no longer wait for the document count; opening a collection, filtering, paging, aggregations and saving an edit are 1.5–8× faster. Load all opens 50,000 documents in a few seconds.
• Lower memory: about 30% less RAM at idle and with a collection open; heavy parts load only when first used.

Fixed

• Selecting rows or opening Bulk Edit no longer blanks the window.
• Schema analysis averages and sample counts are correct; index usage dates are shown.
• A provider picked in AI settings no longer flips back to Gemini.

[1.0.1] — 2026-10-03

Fixed

• Gemini API keys with a dot (newer Google keys look like AQ.…) are accepted, and spaces, line breaks, zero-width characters and quotes picked up when pasting a key are removed.

[1.0.0] — 2026-10-03

First general-availability release of the MotionQL desktop app for macOS (arm64, x64), Windows (x64) and Linux (x64), and of the self-hosted MotionQL Team Server.

Connections and authentication

• Connection Manager with folders, tags, colors and dev/staging/prod environments (production warning banner), quick connect on Cmd/Ctrl+1–9, health indicator, duplicate, and export/import with secrets stripped or encrypted with a password.
• Standalone, seed lists, mongodb+srv, replica sets with read preference and tags, sharded clusters through mongos, direct and load-balanced connections, and Unix domain sockets.
• TLS with CA and client certificate files; SSH tunnels with password or key authentication, host-key pinning and up to four jump hosts; SOCKS5 and HTTP proxies.
• Authentication: SCRAM-SHA-256/1, X.509, LDAP (PLAIN), Kerberos (GSSAPI), AWS IAM and OIDC (browser sign-in with PKCE, and workload identity for Azure, GCP and Kubernetes).

Querying and editing

• Collection tab: filter, projection, sort, skip and limit in shell syntax; collation, read/write concern, read preference, maxTimeMS and hint; visual query builder; cancel; tree, table and JSON views with a column picker; in-place editing; copy as Extended JSON or shell; explain tab; find-to-aggregation conversion; a COLLSCAN / in-memory sort badge on filtered or sorted queries (checked with a queryPlanner explain that does not run the query), linking to the plan and Index Review; document version history with revert.
• IntelliShell: mongosh-style commands parsed without executing JavaScript, with autocomplete, history, snippets, Query Assist and explain; optional Script mode runs JavaScript in a local QuickJS sandbox (can be turned off by policy). Script mode covers the common mongosh helpers, including rs.*, sh.*, db.aggregate, coll.explain(), exact NumberLong math and cursor read preference and read concern; the list of what works is in docs/SHELL_COMPATIBILITY.md. Number literals are typed as in mongosh: whole numbers outside the 32-bit range are stored as Double, not Long.
• Aggregation editor with stage-by-stage preview, full operator library, options and templates.
• SQL Query: read-only SELECT with joins, grouping and ordering, translated to MongoDB queries; .sql files.
• Query Code for mongo shell, Node.js, Python, Java, C#, PHP, Ruby, Go and Rust from deterministic templates.

Data movement and transformation

• Import and export: JSON, JSON Lines, CSV, BSON, Excel and SQL INSERT; column mapping and typing; insert, upsert, replace and merge modes; streamed jobs with progress and cancel.
• Copy collections; dump and restore in mongodump layout with optional gzip, or as one mongodump --archive file that mongorestore reads (and MotionQL restores archives from mongodump, gzipped or not).
• Data Compare & Sync with field-level diffs, dry-run preview, selective sync and conflict policies.
• SQL Migration from PostgreSQL, MySQL, MariaDB, SQL Server and Oracle, with reference/embed relationships, preview and progress.
• Data Model: ER diagrams of a database with fields, keys and relationships inferred from field names, DBRefs, views and matching values; editable layout, relationships, cardinality and notes; export to PNG, SVG, Mermaid and JSON.
• Schema analysis with CSV, Markdown and HTML reports; Reschema operations; Data Masking with eleven masking methods and preview.

Administration and monitoring

• Index management including hidden indexes, usage stats from $indexStats, live build progress (per shard through mongos) and the shard key with the index that backs it; server monitoring; running operations with kill; query profiler; topology view.
• Users and custom roles with a privilege inspector.
• GridFS browser; views; multi-operation transactions with dry-run and review modes; change stream viewer.
• MongoDB Atlas: organizations, projects, clusters, database users and access lists; pause/resume clusters; create connections from clusters.
• Scheduled tasks (export, import, copy, compare, reschema, masking, SQL migration, dump, restore, script, schema snapshot, index review, continuous sync) with once, interval, daily, weekly and cron schedules, missed-run policy and a run log; an optional background runner (launchd, Task Scheduler, systemd user timer) runs them while the app is closed, and a --cli runs them from scripts.
• Dashboards: charts, metrics, tables and notes on a drag-and-resize grid, backed by live aggregations with shared filters, auto refresh and maxTimeMS. A chart editor with a live preview covers column, bar, stacked, line, area, pie, doughnut, scatter, radar and polar-area charts, date grouping (hour to year), series, colour palettes, legends, axis titles, number formats and value labels. Built-in and saved templates, one-click layouts, a present mode, and export to PNG, multi-page PDF (A4 or Letter, cut between widgets), per-widget PNG and CSV, and dashboard JSON.

AI assistant (optional)

• Google Gemini, Anthropic Claude (native, with a model picker), or any OpenAI-compatible endpoint, with the customer's own key (stored encrypted); offline mode.
• Natural language to find, aggregation and SQL; explain plans; fix errors; index advice; pipeline debugging; compare summaries; masking suggestions; schema questions; SQL Migration relationship suggestions.
• Off by default for every connection; sends schema-level context, not documents; suggestions are validated before display.

Security and governance

• Sandboxed, context-isolated renderer with strict CSP; IPC sender verification and allow-listed methods; Electron fuses and integrity-checked app.asar.
• Secrets encrypted with the OS keychain and never exposed to the UI.
• Per-connection read-only, AI and server-side JavaScript policies enforced in the main process, including $out/$merge and explain-wrapped writes; typed-name confirmation for drops.
• Local audit log with credential redaction; app lock with idle timeout, OS-lock and sleep triggers.
• Enterprise policy.json for updates, Team Server URL and sign-in requirement, license deployment, crash-report control and machine-wide security restrictions (read-only hosts, AI, server-side JavaScript, Script mode, background tasks, idle lock), all validated fail-closed.

Team Server (self-hosted)

• OpenID Connect SSO with PKCE (Okta, Microsoft Entra ID, Google Workspace and other OIDC providers) and SCIM 2.0 provisioning.
• Org roles, teams and folder permissions; shared connections (never with secrets), queries, aggregations, shell and SQL scripts, snippets and tasks, with version history.
• Signed effective policies applied by desktops as a floor, including offline.
• Hash-chained, signed and verifiable audit log with desktop event forwarding, export and SIEM forwarding (webhook, syslog over TLS).
• PostgreSQL or SQLite; Docker image running non-root on a read-only filesystem; CLI for keys, audit, users and SCIM tokens.
• Optional read-only web viewer (MOTIONQL_WEB_VIEWER=true) at /web/: shared items with version history, team policy, and the audit log with filters, verification and export, behind the same SSO and roles. Its sessions are read-only on the server.

Platform

• Settings, rebindable keyboard shortcuts and a command palette; light and dark themes; accessibility improvements.
• Auto-update over HTTPS with signature verification, stable and beta channels, no downgrades, and enterprise controls; installers and updates are published to the public motionql-releases repository.
• 14-day trial and offline Ed25519-signed license keys, cancelled through the signed product manifest (revokedLicenses); links to motionql.com to get, renew or upgrade a key; first-run EULA acceptance.
• Opt-in crash reporting and a redacted diagnostics export.
• Signed and notarized builds with SHA-256 checksums, a CycloneDX SBOM and build-provenance attestations; third-party notices.

Known limitations

• No .pkg (macOS) package; no Windows on Arm or Linux arm64 builds.
• IBM Db2, SAP ASE and SQLite are not supported as SQL Migration sources.
• Map-Reduce is intentionally not supported.
• The Team Server web viewer is read-only; administration is through the desktop app, API and CLI.