Name: Monghoul
Author: Monghoul

Query Editor

A full Monaco editor with schema-aware autocomplete over the MongoDB Node driver.

Context-aware completions: field paths, operators, pipeline stages, accumulators, expressions
Progressive nested field disclosure. Type "address." to see child fields only
Pipeline-accumulated fields. $addFields and $group output feeds later stages
$lookup pulls in the foreign collection's schema, nested paths included
Enum value suggestions from schema analysis
JavaScript sandbox with BSON constructors, fetch, crypto, Faker, Luxon, Day.js, Lodash, and Zod
Console and print output lands next to the result
Queries and Markdown notes autosave. Format on demand with Ctrl/Cmd+S

Result Viewer

Six view modes for query results, with inline editing and export.

Table: virtualized spreadsheet with column pinning, sorting, resize, auto-fit, reorder, and multi-select
Tree: hierarchical document viewer with type badges, resizable columns, and sticky document headers
JSON: the whole result as formatted, syntax-highlighted text
Result: a write outcome or single value as a card of copyable metrics
Explain: performance grade, execution plan tree, and index suggestions
Chart (Pro): 8 chart types with dual axis, date aggregation, trend lines, and PNG export
Cross-page search with cell-level match highlighting
Preview drawer (Ctrl/Cmd+I) follows row focus and shows the document as formatted JSON
Export the loaded results or every matching document, as Extended JSON, CSV, or Excel (Pro)

Pro

Charts

Plot a result set without leaving the tab.

Bar, horizontal bar, line, area, pie, and scatter, with stacking for bar and area
Eight guided presets, from time series to correlation to top values
Advanced config: field mapping, several measures, grouping, sorting, legends, colors, dual axes
Aggregate with sum, count, average, minimum, or maximum
Group dates by hour, day, week, month, quarter, or year
Linear trend line on scatter charts
Export as high-resolution PNG with a background that matches the theme

Aggregation Builder

Visual pipeline builder with drag-and-drop, live stage previews, and code sync.

The whole stage catalog in a searchable picker, with stages your server version cannot run marked
Drag-and-drop reordering, and a "+" between every pair of stages
Per-stage enable/disable, collapse, duplicate, clear, and "Run to here"
Index-aware autocomplete the main editor does not have: indexed fields rank first in $match and $sort
A collapsed stage prints its body on one line, so a folded pipeline still reads
Live previews with an optional 20-document safety limit, cancellable from the footer
Errors name the root cause instead of the server's wrapper chain
$lookup form helper, synced both ways with the stage body
$out and $merge are marked destructive and honor write protection
Undo/redo up to 50 entries, and code that stays in step with the visual pipeline

Explain View

Understand query performance at a glance, for both find and aggregation queries.

Performance grade with execution time, docs returned and examined, keys examined, scan type, and index name
Per-stage row counts and elapsed time across the whole pipeline, sub-pipelines included
Data flow funnel showing document counts shrinking through each stage
Index suggestions ordered by the ESR rule, extending your current index instead of replacing it
It skips the noise: synthesized fields, identity indexes, and indexes already in place
Warnings for collection scans, in-memory sorts, disk spills, and misplaced limits
Flags an Atlas $search or $vectorSearch stage that is not first in the pipeline
Auto-explain after every read query. On by default, switchable off per connection

Try the Explain Visualizer online, no install needed →

Connection Management

Connect to any MongoDB instance with full protocol support.

8 auth methods: no auth, username/password, SCRAM-SHA-1/256, and X.509, LDAP, Kerberos, AWS IAM on Pro
Turn off "save password" and it never touches disk. Monghoul asks once per app session
SSH tunneling with password or private key authentication
SSL/TLS with a custom CA certificate and an option to allow invalid certificates
Write protection over writes, drops, renames, index changes, and $out/$merge, scoped per database or collection
Database and collection filters to tidy the tree. They are not a permission boundary
Connectivity and credentials are tested before a connection saves

Inline Editing

Edit document values directly in the table or tree view without opening a separate editor.

Double-click any cell to edit: strings, numbers, booleans, dates, ObjectIds, null
Type selector dropdown to change the field type where the conversion is valid
Date picker with calendar, time spinners, quick presets, and time zones
Enum combobox suggests known values from schema analysis while allowing custom input
Write protection confirmation on protected connections before saving
A value the chosen type cannot hold is refused with the limit named. Nothing is written
Optimistic updates. The edited value appears immediately and rolls back on failure

Schema Analysis

Sample-based schema analysis with enum detection and editor integration.

Configurable sample size with presets. Free analyzes up to 500 documents per run
Hierarchical field tree with type badges, array element types, and occurrence bars
Automatic enum detection with configurable thresholds, plus field paths you can force
Manual enum editing, and enum values feed autocomplete in the query editor
Automatic analysis is global, and each connection can override it on or off
Collections a query reaches through $lookup or $unionWith get analyzed too
Schema banner generates a missing schema in one click, right from the editor
Field search, cancellable runs, and results cached across sessions

Pro

Cluster Monitoring

Live server metrics, trend charts, current operations, and the database profiler.

Connections, op and document rates, memory and cache, network, locks, cursors, and scan efficiency
Replica-set member health, latency, and lag
Health strip rating cache, connections, replica lag, lock queue, and active ops as healthy, warning, or critical
Sparkline trend charts with time-aware hover values
Live operations with duration and namespace filters, scan warnings, and confirmed kill
History keeps 200 operations of one second or longer for the tab session
Profiler set to off, slow-only, or all, with collection summaries and JSON/CSV export
Refresh every 1, 2, 5, 10, 30, or 60 seconds, with pause and resume

Pro

Users & Roles

Manage MongoDB users and custom roles, on self-hosted deployments and on Atlas.

List, filter, create, edit, and delete users. Generate, reveal, and copy a secure password
Custom roles with resource-scoped privileges, grouped actions, and inherited roles
Role detail pane: flattened privileges marked granted or inherited, and who holds them
The create forms show the equivalent mongosh command, with passwordPrompt() in place of the password
Atlas connects with a project API key or a service account. The secret is stored encrypted on the device
On Atlas the forms adapt: cluster scopes, temporary-user expiry, and only the actions Atlas can grant
Deleting the last user administrator warns first. Mutations are logged, passwords never are

Pro

MCP Server (AI Tools)

70+ Model Context Protocol tools so an AI assistant can drive the app.

Ten essential tools are always on, from list_connections and execute_query to create_query_tab
The client switches the rest on itself with discover_tools, enable_tools, and disable_tools
Capability groups for queries, schema, indexes, import/export, tabs, panels, windows, charts, and logs
Configurable port and a regenerable auth token, with snippets for the common AI clients
Review mode holds an AI-initiated query until you approve it
Access can be limited to chosen connections and databases. Anything else is hidden and refused
Access log keeps the 100 most recent calls. Tool parameters are not recorded

Import & Export

Move data in and out of MongoDB in any format.

Import JSON, Extended JSON, gzipped JSON, CSV, and TSV. NDJSON and Excel on Pro
Picking a file sets the format, suggests the collection name, and fixes the delimiter
Write strategies: insert, upsert by identity, or drop then insert
Type inference for numbers, booleans, null, and JSON cells. Dates only from a leading YYYY-MM-DD
Export Extended JSON (canonical or relaxed, array or NDJSON, gzip), CSV, or Excel (Pro)
CSV and Excel columns come from a sample of the export, not just its first document
Copy a collection to another database or connection, with optional index replication (Pro)
Import or export a whole database, one file per collection (Pro)
Progress and cancellation from the footer on every long transfer

Index Management

Create, inspect, and manage indexes with usage statistics and index suggestions.

Create compound indexes with ordered fields and an ascending/descending toggle
Options for name, unique, sparse, and a partial filter expression
The exact createIndex() command is previewed before creation
Explain recommendations prefill the Create Index form in one click
Visual indicators for unique, sparse, partial, and TTL indexes
Per-index size and usage count from $indexStats, color-coded by activity
Drop or rebuild individual indexes, and refresh statistics from the server

Workspace & Tabs

Multi-panel layout, detached windows, and a command palette.

Query, Markdown, and JSON Viewer tabs, plus Cluster Monitor and User Management on Pro
Markdown tabs with a split view, formatting toolbar, GitHub-flavored Markdown, and autosave
Drag-and-drop tab reordering within and across panels, with full state carried along
Split panels horizontally or vertically with nested splits and named panels
The 100 most recently closed tabs can be restored, newest first, with their full state
Leaving a tab does not cancel its query. Background work stays cancellable from the footer
Detach any tab into its own window and reattach it later (Pro)
Command palette (Ctrl/Cmd+K) with fuzzy search across everything

Sidebar Organization

Folders, pinned collections, favorites, and drag-to-reorder across all sections.

Create folders and nested subfolders in Connections, Pinned, and Favorites
Drag items into folders or back to root with VSCode-style three-zone drop detection
Collapse/expand folders with item count badges. State persists across sessions
Pinned collections open a pre-formatted find query on double-click
Customizable snippets. Edit templates, reorder, and set the default query
Favorites saved with Ctrl/Cmd+Shift+S, keeping view type, chart config, and page size
Markdown notes can be saved as favorites too
Hover preview cards for connections, databases, collections, indexes, and schemas

Data Generation

Generate synthetic documents with Faker for testing and development.

100+ Faker providers across 16 categories: names, emails, addresses, commerce, finance, and more
Value modes: Faker, literal, list, or numeric range, with per-field null probability
Nested objects and configurable arrays
Seed the tree from a real collection and it keeps each field's type, Decimal128 included
A field under a unique index gets no repeats, and GeoJSON coordinates come out valid
Preview a sample document before generating
Up to 100,000 documents per run on Pro, 500 on Free

Operation Logs

Full audit trail of every operation, doubling as a past queries search tool.

Search query text, connection, database, collection, and error messages
Filters for operation type, status, connection, collection, and time range
Compare two query logs side by side, or reopen one as a query tab
Retention of 100 to unlimited entries, 250 on Free
Response storage of none, stripped (the default), or full, capped between 5 KB and 1 MB
Full storage keeps the response unredacted on this device, and says so before you pick it
Export the filtered set, or export server logs for a support report

Themes

11 built-in themes and a theme editor that derives a full palette from three colors.

Monghoul Dark (default), Monghoul Light, Midnight Ocean, Rosé Pine, Nord Aurora, Ember Glow
Catppuccin Mocha, Sandstone, Arctic Light, Ink, and Paper
Every built-in theme meets WCAG AA contrast: 4.5:1 for text, 3:1 for chart series and other indicators
Chart series stay distinguishable from each other within a theme
Build a custom light or dark theme from three colors, previewed live (Pro)
Custom themes reach editor syntax, MongoDB value types, result views, and index badges (Pro)
Import and export themes as JSON (Pro)

Pro

Document Diff

Compare any two documents side by side with line-level change highlighting.

Pick any two from a selection of two or more documents
JSON formatting with syntax highlighting on both sides
Swap button to flip left and right
Configurable label field for the document dropdowns

Date Helper

Build date expressions visually instead of typing ISODate strings by hand.

Single date or range mode, with independently selectable endpoints
Presets evaluated in the selected time zone: today, this week, last month, last 7 days, and more
Named time zones and UTC offsets
Shortcuts for the current moment, midnight, a whole day, and now as a range end
Copies the expression to the clipboard as either new Date or ISODate

Validation Rules

View, create, and edit MongoDB collection validators.

Monaco JSON editor for the validator expression
$jsonSchema scaffold insertion with one click
Validation level (off, moderate, strict) and action (error, warn) controls
Clearing the validator and saving removes the rules, with a confirmation
Validation node in the sidebar tree with level and action badges

Complete Feature Reference

Every capability in Monghoul, grouped by area. Anything marked Pro needs a Pro license or the 14-day trial.

Connections & Security

8 authentication methods: no auth, username/password, SCRAM-SHA-1, SCRAM-SHA-256, and on Pro X.509, LDAP, Kerberos, and AWS IAM
Configurable authentication source for any of them
Connect from a URI or from individual fields. Pasting a URI fills in host, port, credentials, auth, TLS, replica set, and read preference
Both mongodb:// and mongodb+srv:// are supported
Save password can be turned off for username/password and SCRAM. The password is then written nowhere on disk, and Monghoul asks for it once per app session
SSL/TLS with a custom CA certificate and an option to allow invalid certificates for self-signed setups
SSH tunneling with password or private key file, and a tunnel that closes with its connection
Direct or replica-set topology with read preference: primary, primaryPreferred, secondary, secondaryPreferred, nearest
Per-connection result cap, default 1,000 documents. An explicit .limit() wins over the cap even when it is larger
Auto-explain per connection. Read queries collect plan statistics automatically
Automatic query execution the first time a tab is opened in a session, per connection
Cancellable queries. Stopping one, by hand or by timeout, also ends the operation on the server
Configurable connection timeout, default 30 seconds
Database and collection filtering in ignore (hide listed) or allow (show only listed) mode. System databases are filtered by default
Those filters are organizational, not a security boundary. They do not change MongoDB permissions or stop raw query code from naming a hidden collection
Write protection scoped to a connection, or to selected databases and collections. It covers inserts, updates, deletes, bulk writes, drops, renames, index changes, and aggregation output stages
Protected operations require explicit confirmation, at every entry point including the keyboard
Protected connections, databases, and collections are marked in navigation and previews
Connectivity and credentials are tested before a new connection saves, and any connection can be re-tested on demand
Unique connection names, generated from host, port, and database when you do not supply one
Connection color coding with 12 presets and a custom picker. The color propagates to tabs and sidebar entries
Deleting a connection closes its tabs and removes its pins and favorites. Its log history is kept

Sidebar & Navigation

Hierarchical tree: Connections > Databases > Collections > Indexes, Schema, Validation, with lazy-loading expansion
Collections sorted alphabetically within each database
Instant search over the whole hierarchy. Matches auto-expand, the field reports how many matched, and clearing it restores the previous expansion state
One click focuses a collection, a double-click opens it in a query tab
Refresh pulls a collection with its statistics and indexes. With nothing focused, the active query tab is the target
Hover preview cards on connections: host with credentials masked, database and collection counts, data size, topology, auth method, transport security, and write-protection scope
Hover preview cards on collections: document count, data size with average per doc, schema status, index count and size, and validator info
Hover preview cards on databases, indexes, and schema nodes with their own relevant detail
Previews follow the keyboard cursor as well as the pointer
Right-click context menus at every level, from create database and refresh down to rename, clear, and drop
Pro context actions: Cluster Monitor, user management, database import, and database export
Customizable snippets submenu. Add, remove, drag-to-reorder, and edit query templates, then star one as the default double-click query
Pinned collections in a dedicated section, keeping the same context menu and hover preview as regular collections
Favorite queries saved with Ctrl+Shift+S / Cmd+Shift+S. Preserves code, view type, chart config, table state, page size, and read preference
Markdown notes can be saved as favorites too, and reopen as markdown tabs
Folders for all three sections, nested to any depth, with rename in place and item count badges
Deleting a folder returns its contents to the section root rather than deleting them
Drag-to-reorder connections, favorites, and folders with three-zone drop detection
Drag a collection or database onto another destination to open a copy modal with it preselected (Pro)
Sticky headers keep connection and database rows visible while scrolling their descendants
The active tab's collection is marked in the tree, and a disconnected connection is dimmed
Pin and favorite changes appear in every open Monghoul window
Resizable sidebar (5-80%) with width, order, and expansion state persisted. Toggle with Ctrl+B / Cmd+B
Reaching a Free capacity limit explains the Pro upgrade instead of failing quietly

Command Palette

Open with Ctrl+K / Cmd+K from anywhere in the app
Fuzzy search over open tabs, recently closed tabs, connections, databases, collections, and favorites
Pinned collections rank above unpinned ones
Tab and favorite search covers query code as well as titles, and shows the matching excerpt
Filter results by category, with a live count per category
With no search term it shows recent and relevant items
Choosing an item switches to it, opens it, restores it, or reveals it in the sidebar
Full keyboard navigation, with the selected action named before it runs

Query Editor & Environment

Full Monaco (VS Code) editor with JavaScript syntax highlighting, configurable font size, and theme-aware styling
Inline syntax and type errors, checked against the query environment's own declarations
The environment is the MongoDB Node driver, not the mongo shell. A projection passed as find's second argument is refused rather than quietly ignored
db.orders is shorthand for db.collection('orders'), and it yields to real driver methods, so stats, admin, and command need the explicit form
A single expression returns its value automatically. A multi-statement script runs in order and returns an explicit result
A query is not reported finished until the database operations it started have finished
Sandbox globals: db, ObjectId, the id() shorthand, ISODate, NumberLong/Int/Decimal, Extended JSON helpers, print, sleep, URL, encoding, crypto, and fetch
Bundled libraries: Faker, Luxon, Day.js, Lodash, Lodash FP, and Zod
BSON numeric constructors keep their type and precision. A value the type cannot hold is rejected, not truncated
Console and print output appears with the result, warnings emphasized, capped at 500 entries per run
Code formatting with Prettier via Ctrl/Cmd+S, a floating button, or Shift+Alt+F inside the editor
Method chains break onto separate lines without changing behavior, and generated code is formatted on tab creation
Run with Ctrl+Enter / Cmd+Enter / Cmd+R, plus Explain and Stop buttons and an elapsed timer
Cancellation stays available while the connection is still being established
Result timing separates query, connection, server, and total duration
Changing a tab's connection keeps the selected database only when the new connection has one by that name
Save to, sync with, or remove from favorites straight from the toolbar. Queries and notes autosave

Schema-Aware Autocomplete

Field path suggestions with type annotations (String, Number, ObjectId, Date) derived from schema analysis
Progressive nested field disclosure shows only the current nesting level
Pipeline-accumulated fields. $addFields, $group, $project, and $lookup output feeds autocomplete in later stages
Reshaping stages replace the field list, removal stages filter out dropped fields, and disabled stages are skipped
$lookup resolves the foreign collection's schema for nested paths like customer.address.city
All MongoDB query operators, update operators, and 40+ collection and cursor methods with inline documentation
All aggregation stages, accumulators, and 50+ expression operators
Stage-aware string completions. $lookup.from suggests collections, $unwind suggests array fields, $group._id suggests field references
$facet branches expose sub-pipeline output, and $unionWith merges the other collection's schema
Enum value suggestions inside $in, $nin, $eq, $ne, and direct field comparisons
Type-aware operators: $regex for strings, $gt/$lt for numbers and dates, $size/$elemMatch for arrays
$expr context switching. Inside $expr in $match, autocomplete moves to aggregation expressions
System variables ($$ROOT, $$NOW, $$REMOVE) and variables from $let, $map, $filter, and $reduce, in scope only
BSON constructors with snippet placeholders: ObjectId(), NumberLong(), NumberDecimal(), UUID(), BinData(), Timestamp()
$jsonSchema keyword suggestions inside validator expressions
Suggestions are advisory. They never block execution, and stale schema data degrades them without breaking MongoDB language completions
Index-aware ranking is an aggregation builder feature. The main editor does not have it

Aggregation Builder

Visual builder with the whole stage catalog ($match, $group, $lookup, $facet, $graphLookup, $setWindowFields, $densify, $search, $vectorSearch, and more) in a searchable, categorized picker
Server version gating. Stages your server cannot run are marked and unavailable; the raw editor stays ungated
Color-coded stage badges by category: Filter & Sort, Transform, Group & Aggregate, Join & Combine, Output, Other
New stages start with useful templates, updated when you change the operator
Quick-start templates for empty pipelines: Filter & Sort, Group & Count, Join Collections, Reshape Fields
Insert a stage anywhere from the "+" on the connector between every pair
Reorder by drag-and-drop or with Move Up / Move Down
Per-stage enable/disable, collapse, duplicate, clear, delete, and Run To Here
Expand, collapse, enable, disable, or clear every stage in the pipeline in one action
A collapsed stage prints its own body on one line, so a folded pipeline can be read without expanding it
Syntax errors are caught before execution and shown on the stage itself, expanded or collapsed
Schema-, type-, index-, variable-, and pipeline-aware autocomplete per stage editor
Index awareness the main query editor does not have: indexed fields and valid compound-index continuations rank first in $match and $sort
$lookup form helper with From, Local/Foreign field, and As dropdowns, synced both ways with the stage body
Explain plan button that runs on the current pipeline without modifying the code
Stage previews inline, on collapsed cards as well as expanded ones, with an optional 20-document safety limit
Auto-preview mode re-runs after each edit, and previews are cancellable from the current-operations panel
An edit discards previews from that stage downward, so a card never shows an older pipeline's result as current
A failed preview names its root cause rather than the server's wrapper chain, with the original message copyable
$out and $merge are visibly marked destructive and honor connection write protection
Stage search appears once a pipeline has four or more stages. Filter by operator name, description, or body
Undo/redo over structural changes, up to 50 history entries
Visual and code modes round-trip. The generated code sits in a togglable side panel and is copyable in place
Pipeline structure, enabled and collapsed state, collection, and editor mode persist with the tab

Result Views

Six view modes: Tree, Table, JSON (the result as formatted text), Result (a write outcome or single value as a card of copyable metrics), Explain, and Chart (Pro)
Active mode persists per tab. Set Tree or Table as the default for new queries, or restore automatic selection
Fullscreen: Tab (results fill the tab) or Window (results fill the app), exit with Escape
Collapsible result panel and a resizable editor/result split, both persisted per tab
Preview drawer (Ctrl/Cmd+I) follows row focus in Tree and Table, showing the document as formatted JSON
Tree view: keyboard expand and collapse, MongoDB-aware type badges, resizable columns, and sticky document headers you can switch off per tab
Table view: column sorting, resize, auto-fit, pinning, and reorder, persisted per tab and in favorites
Headers identify the predominant column type
Ctrl/Cmd+F searches the loaded result set, highlighting individual matches in keys, cells, and nested values
Crossing a page boundary moves to the next loaded match. It does not issue a server-side search or fetch past the result cap
Hover preview for Object and Array values, with formatted JSON after 400ms
Expand complex values inline with a nested tree, breadcrumb path, editable leaves, and a resizable panel
Result cap with cursor pagination. Load 5K / 10K / All when capped, offering only the choices above the current cap
Loading everything warns first that a large collection may be slow
Auto-explain badge on every result: green IXSCAN with index name, or red COLLSCAN, with the full plan on hover
Empty result sets still show timing and explain information
Pagination with previous, next, and direct page entry. Page sizes 10 to 1,000, or any value from 1 to 10,000

Document Actions & Result Export

Edit Document hotkey (Ctrl/Cmd+E) on the focused Tree or Table row. No right-click required
Inline editing for strings, numbers, booleans, dates, ObjectIds, and null, with a type selector and enum combobox
Write protection confirmation before an inline change, and optimistic updates that roll back on failure
A value the chosen type cannot hold is refused with the limit named, and nothing is written
Right-click menu: Edit Value, Filter by Value, Unset Field, Clone, Copy Value/Key/Path/Column/Document, Edit, Delete, Insert
Tabs generated from a result open next to their source tab
Multi-select by item, range, or whole page, with a floating bar for Copy, Delete, Show Diff, and Cancel
Bulk delete with a confirmation listing the identifiers it will remove, or open it as an editable query instead
Deleted rows leave the result immediately and come back if the server rejects the delete
Document diff on any two of a 2+ row selection, with a swap button and configurable label field (Pro)
Export in two steps, scope then format: the loaded results or every matching document, then Extended JSON, CSV, or Excel (Pro)
The all-matching scope re-runs the query with .limit(), .skip(), and the result cap dropped, keeping limits inside nested pipelines
It warns that the read is unbounded and asks for confirmation, and asks again when the query could mutate data
All-matching Extended JSON writes one document per line; a loaded-results export writes one indented array
A result that cannot be re-read as documents says why the all-matching scope is unavailable

Charts & Visualization (Pro)

8 chart types: Bar, Horizontal Bar, Line, Area, Pie, Scatter, Stacked Bar, and Stacked Area
Eight guided presets: Time Series, Category Breakdown, Distribution, Multi-Series, Trend Analysis, Correlation, Stacked, and Top Values
Advanced mode with manual X/Y fields, several Y-axis series, grouping, and the full option set
Aggregation methods: sum, count, average, minimum, maximum
Date aggregation grouping by hour, day, week, month, quarter, or year
Dual Y-axis for comparing values at different scales
Customizable title, legend, labels, sort order, data limit, series colors, stacking, and filters
Chart theming follows the active app theme, with series guaranteed distinguishable within it
Fullscreen (Tab and Window) consistent with every other view type
Export as high-resolution PNG (4x pixel ratio) with a theme-appropriate background
Chart configurations saved per tab and persisted with favorites

Tabs & Workspace

Tab types: Query, Markdown, and JSON Viewer (read-only tree), plus Cluster Monitor and User Management on Pro
Markdown tabs with editing, a side-by-side split, or preview, a formatting toolbar, GitHub-flavored Markdown, task lists, and sanitized embedded HTML
Markdown tabs autosave, and save to favorites the way query tabs do
An empty markdown tab opens in the editor; one with content opens in preview
Drag-and-drop tab reordering within a panel and across panels
Tab context menu: Execute/Stop Query, Clone, Reveal in Sidebar, Rename, Split Right/Down, Move to Panel, Open in New Window, Close, Close Others
Inline rename by double-clicking the title, a connection color indicator per tab, and middle-click to close
Restore up to the 100 most recently closed tabs, newest first, with complete state
Moving a tab between panels preserves its editor, result, pagination, preview, and toolbar state
Leaving a tab does not cancel a running query or a background operation
Background work stays cancellable from the footer, and inactive tabs pause their own polling and timers
Multi-panel layout: horizontal, vertical, and nested splits, resizable dividers, and custom panel names
Detach any tab into a floating window and reattach it later, with size and position persisted (Pro)
Main window bounds saved and restored on launch

Import & Export

Import JSON (array, Extended JSON, gzip) and CSV/TSV with configurable delimiter, header handling, and per-column type mapping
NDJSON and Excel (.xlsx/.xls) import on Pro. An Excel import reads the first sheet and takes row one as the column names
Choosing a file sets the format from its extension, suggests the collection name, and picks the tab delimiter for a .tsv
Write strategies: Insert, Upsert by identity, or Drop & Insert
Drop & Insert into a write-protected target requires confirmation before the drop
Stop on the first error or skip invalid records and continue
Type inference for numbers, booleans, null, and cells holding a JSON object or array. An empty cell imports as null
Dates only from an ISO-style leading YYYY-MM-DD, so order-1, sku-2, and Q1-2024 stay text
Import an entire database from a directory, one collection per file, with the file count reported up front (Pro)
Export Extended JSON (canonical or relaxed, array or NDJSON, optional formatting and gzip) or CSV (custom delimiter, array flattening, gzip)
Excel export on Pro, with Extended JSON aware serialization
CSV and Excel columns are the union of fields in a sample of the export, so a field only some documents carry still gets one
A field that first appears after that sample gets no column, and the export reports which fields those were
Export scope narrowed by a document limit and a MongoDB filter
Copy a collection to another database or connection, with Insert, Upsert, or Insert Only (Pro)
Collection copy can carry non-identity indexes and preserve their supported options (Pro)
Export or duplicate a whole database with indexes, over every collection or ones picked from a searchable list (Pro)
A database export writes one Extended JSON or CSV file per collection into a directory you choose (Pro)
Long transfers report progress and cancel from the footer. A cancelled export removes its file; a cancelled import or copy keeps what it wrote

Schema Analysis

Sample a configurable number of documents, with presets for the common sizes. Free analyzes up to 500 per run
Real-time progress with cancellation, listed in the footer under its collection name
Hierarchical field tree covering nested objects and arrays
Color-coded type badges for every detected BSON type, with arrays showing element types like array<number|string>
Per-field occurrence bar. Fields present in 100% of the sample are marked required
Stats summary: total fields, required, optional, enum count, sample size, and analysis time
Automatic enum detection with configurable thresholds: max distinct values (default 20) and max ratio (default 0.2)
Optional numeric enum support, and named field paths you can force to count as enums regardless
Manual enum editing. Add values, hover a chip to remove it, and edits persist immediately
Enum values drive autocomplete inside $in, $nin, $eq, $ne, and direct comparisons
Field search filters the tree by name and auto-expands the parents it needs
Automatic analysis and sample size are global settings, and each connection can inherit or override them
Per-connection overrides force automatic analysis on or off and set the sample to 50, 100, or 500 documents
Collections reached through $lookup.from, $graphLookup.from, $unionWith, $merge.into, or $out are analyzed automatically
Schema banner at the bottom of the editor generates a missing schema in one click
A collection with no documents says it has nothing to analyze, rather than reporting a missing schema
Copy the full schema as JSON, and results cache across sessions with a relative last-analyzed timestamp

Data Generation

Visual schema builder for the generated document structure
Supported types: String, Number, Boolean, Date, ObjectId, Object (nested), Array (configurable length and element type)
100+ Faker providers across 16 categories: names, emails, addresses, phone numbers, companies, commerce, finance, lorem, dates, images, vehicles, and more
Four value modes per field: Faker, Literal, List (random from a set), and Range (numeric min/max with decimal places)
Per-field null probability from 0 to 100% for realistic sparse data
Seed the tree from a sampled collection and it keeps each field's type: array element types, Long, Int32, Double, Decimal128, and fields only sometimes null
A field a unique index covers is generated so no two documents in a run repeat a value
A GeoJSON coordinates field is generated as valid pairs, so a 2dsphere index accepts it
Preview and copy a sample document before generating
Up to 100,000 documents per run on Pro, 500 on Free, with background progress

Index Management

Create multi-field compound indexes with reorder buttons and an ascending/descending toggle
Index options: custom name, unique, sparse, and partial filter expression (JSON)
The exact createIndex() command is previewed before creation
Explain suggestions open the Create Index form with fields already ordered by the ESR rule
Per-index size and usage statistics from $indexStats, color-coded green (active) or muted (unused)
Usage tracking start time alongside the operation count
Visual indicators for unique, sparse, partial filter, and TTL indexes, with hover tooltips
Index details: key fields with direction, flags, options, and size
Drop or rebuild individual indexes, and refresh definitions and statistics from the server
Indexes folder hover preview with total size, total access operations, and unique/sparse counts

Validation Rules

Full Monaco JSON editor for the validator expression with syntax highlighting
One-click $jsonSchema scaffold insertion
Validation level: Off, Moderate (existing documents exempt), or Strict (all documents validated)
Validation action: Error (reject invalid documents) or Warn (allow with a warning)
Clearing the validator and saving removes the rules, with the same confirmation Remove asks for
Sidebar tree node for collections with validators, showing level and action in a badge
Refresh validator metadata from the server to pick up changes made outside the app

Cluster Monitoring (Pro)

Live Cluster Monitor tab per connection, with host, MongoDB version, storage engine, uptime, and replica-set identity
Metrics for connections and capacity, operation and document rates, memory and cache, network traffic, locks and queued work, cursors, and scan efficiency
Replica-set member health, latency, and lag
Health strip rating cache use, connection use, replica lag, lock queue, and active operations as healthy, warning, or critical
Short trend charts with time-aware hover values on metrics that have history
Operations view with Live, History, and Profiler modes
Live mode: duration and namespace filters, sortable data, scan and lock warnings, command details, and confirmed termination
History retains up to 200 operations of one second or longer for the tab session
Profiler sets a database to off, slow-only, or all, and reports the threshold in effect
Profiler mode shows the 200 most recent profiled operations, with collection summaries and JSON/CSV export
Refresh interval of 1, 2, 5, 10, 30, or 60 seconds, with pause and resume
Permission errors explain the MongoDB privileges the metric requires
Monitoring stops after an error and resumes only when you retry

Users & Roles (Pro)

Connection-scoped workspace for users and roles, using the connected account's effective permissions
Users: list, filter, create, edit roles, change password, and delete
Password users can generate, reveal, and copy a secure password
Identities from normal authentication databases and from external authentication
Role assignment distinguishes built-in from custom roles and shows cross-database scope
Built-in roles are read-only. Custom roles can be created, edited, and deleted
Custom roles support resource-scoped privileges, grouped action selection, and inherited roles
An invalid cross-database or cluster privilege scope is refused, with the offending scope named
Role detail pane: flattened effective privileges marked granted or inherited, the roles it inherits, and which users and roles hold it
Create User and Create Role show the equivalent mongosh command, collapsed and copyable, with passwordPrompt() instead of the typed password
Atlas connects with a project API key or a service account authenticated by OAuth2
The secret half is stored encrypted on this device and can be disconnected at any time
A service account secret shows its expiry date and warns within 30 days of it
On Atlas the user form adds a description, temporary-user expiry, and scopes for the clusters the user may reach
On Atlas the role form offers only the actions Atlas can grant, and not anyResource
The workspace adapts to available access: full administration, view-only, Atlas read-only until credentials are connected, or Atlas managed
A deployment with no access control works too, with a warning
Deleting the last user administrator warns that it appears to be the last one
Every user and role mutation is recorded in operation logs. Passwords never are

Operation Logs

Audit trail and query history over queries, imports, exports, copies, renames, creates, deletes, and other mutations
Each record holds the operation, its connection, database and collection context, the query, duration, status, error, and result metadata
Search across query text, connection, database, collection, and errors, matching any part of a value
Filters for operation type, status, connection, collection, and time range
The collection filter is picked from a list of names and matches exactly
Keyboard navigation, single and multi-select, bulk delete, and export of the filtered set
Compare two query logs side by side in a Monaco DiffEditor with line-level highlighting
Reopen a log as a query tab when its connection still exists
Query, response, error, and metadata sections copy and collapse independently, with full responses fetched on demand
Response storage: none, stripped with representative content (the default), or full by explicit choice
Any stored response is truncated above a size cap you set between 5 KB and 1 MB
Full storage keeps sensitive data on this device, so choosing it carries a warning
Passwords entered in Monghoul's own forms are never logged, but query text and response fields are not classified or redacted
Retention of 100, 250, 500, 1,000, 2,000, 5,000, or unlimited entries, 250 on Free, applied immediately
Approximate log storage shown, with a warning above 50 MB. The list paginates at 50 per page
Export server logs, or attach app version, license tier, platform, and recent logs to an issue report

Current Operations

One panel lists every operation the app runs against a server, from imports and exports to previews, connection tests, and inline edits
Imports, exports, copies, generation, and requested schema analysis keep running whether their modal is open or not
Each row names the operation and its connection, with a progress percentage where the work can be measured
Operations started in another window, or by the AI assistant, appear here too
Cancellation stops the work, not just its progress bar
A cancelled export removes its incomplete file. A cancelled import, copy, or generation keeps what it wrote
Only running operations can be cancelled. A finished one stays listed briefly with its outcome
The footer shows how many operations are listed and aggregate progress while any are running

MCP Server (AI Integration, Pro)

70+ Model Context Protocol tools. Works with Claude, Cursor, Windsurf, and any MCP-compatible client
Enable or disable the server, choose its port, and generate or replace its authentication token
Ready-to-copy connection snippets for the common AI clients
Ten essential tools always available: list_connections, sync_databases, sync_collections, execute_query, get_tab_result, create_query_tab, get_workspace, update_tab_code, activate_tab, show_toast
The discover_tools, enable_tools, and disable_tools meta-tools are always available too
The client discovers and toggles the optional capability groups itself. Essential and meta-tools cannot be switched off
Capability groups for connections, queries, schema, statistics, import/export/copy, indexes, tabs, panels, windows, sidebar organization, favorites, pins, themes, charts, notifications, logs, background tasks, and replica-set status
Review mode queues AI-initiated queries for manual approval before they run
Access can be unrestricted or limited to selected connections and databases
Under restriction, inaccessible items are hidden from discovery and any action targeting one is refused
Connection-wide access includes databases discovered later. Database-specific access does not
MCP access limits organize what the assistant reaches. They do not replace MongoDB permissions
AI-initiated query and chart changes appear in the open app without a manual refresh
Access log keeps the 100 most recent entries with tool name, outcome, duration, and timestamp
Tool parameters are not recorded, and the log can be viewed or cleared
Expired Pro access disables the server

Themes & Appearance

11 built-in themes: Monghoul Dark (default), Monghoul Light, Midnight Ocean, Rosé Pine, Nord Aurora, Ember Glow, Catppuccin Mocha, Sandstone, Arctic Light, Ink, and Paper
Every built-in theme meets WCAG AA contrast on every surface it paints: 4.5:1 for text, 3:1 for non-text indicators such as chart series
Chart series are guaranteed distinguishable from one another within a theme
A brand accent for tier badges and highlights, kept distinct from the primary action and status colors
Theme changes apply instantly across the app, with no restart
Theme editor with 3-color generation. Pick Brand, Background, and Foreground to derive 40+ semantic tokens, previewed live (Pro)
Custom themes cover app colors, semantic and interaction states, MongoDB value types, typography, editor syntax, result views, and index badges (Pro)
22 MongoDB type-specific colors for data type badges, customizable per theme (Pro)
Monaco editor theming with custom token rules (Pro)
Import and export themes as JSON files to share (Pro)

Interface & Display

Hide the footer items, sidebar sections, result views, result-header actions, query-toolbar actions, and secondary context-menu actions you do not use
The result header can go whole or item by item
The connection tree and app-menu entry always stay available
At least one non-chart result view must stay enabled, and hiding the active one selects another
Hiding an action never removes an equivalent keyboard or command-palette path
UI zoom from 50% to 200%, with reset and Ctrl/Cmd+scroll
The status bar shows the current zoom and opens a control with step buttons, presets, and a reset that names its shortcut
Result dates display as ISO, local, relative, or Unix
Contextual tips can be turned on or off
Automatic updates, telemetry, automatic schema analysis, and schema sample size are configurable
Visibility and zoom settings persist per installation

Keyboard Shortcuts

Ctrl+Enter / Cmd+Enter: run query (or Cmd+R)
Ctrl+K / Cmd+K: open command palette
Ctrl+N / Cmd+N: open the add-connection dialog
Ctrl+S / Cmd+S: format the active query editor. Shift+Alt+F also formats, from inside the editor
Ctrl+Shift+S / Cmd+Shift+S: save or update a favorite
Ctrl+B / Cmd+B: toggle sidebar visibility
Ctrl+W / Cmd+W: close active tab
Ctrl+Shift+T / Cmd+Shift+T: restore last closed tab
Ctrl+D / Cmd+D: clone active tab
Ctrl+1-9 / Cmd+1-9: switch to tab by position
Ctrl+Tab / Ctrl+Shift+Tab: next/previous tab
Ctrl+Alt+Arrow / Cmd+Option+Arrow: next/previous panel
Ctrl+Plus / Ctrl+Minus / Ctrl+0, numeric keypad included: zoom in, out, reset
Ctrl+F / Cmd+F: search the focused sidebar or result view
Ctrl+E / Cmd+E: edit the focused document. In a markdown tab, Ctrl/Cmd+E and Ctrl/Cmd+P switch editor and preview
Ctrl+C / Cmd+C: copy the focused cell or selected rows
Ctrl+A / Cmd+A: select every row in the focused result view
Ctrl+I / Cmd+I: toggle the preview drawer
Ctrl+Shift+R / Cmd+Shift+R: refresh the focused or active collection
Escape: cancel or close the active transient UI
Most global shortcuts can be searched and rebound, with conflicts confirmed before reassignment
Rebinding replaces all of a command's defaults with the single new combination
Reset one binding or all of them. Changes take effect immediately and persist per installation
Escape, tab-index switching, the zoom scroll gesture, Shift+Alt+F, and in-view navigation keys are fixed
Shortcut labels adapt to the operating system, and the active UI context decides which command receives a shared binding
Result-view shortcuts work in both Table and Tree, and are rebindable in either

Extended JSON & BSON

ObjectId, Date, Binary, Timestamp, UUID, MinKey, MaxKey, RegExp, Code, DBRef, and the numeric BSON types all get type-aware presentation
Color-coded type badges stay consistent across Tree View, Table View, inline editors, and export output
Shell-compatible constructors in the editor: ObjectId(), ISODate(), NumberLong(), NumberDecimal(), UUID(), BinData(), Timestamp()
Copy, edit, insert, clone, and filter actions generate runnable constructor syntax when the decoded result kept its type and exact value
Results display as relaxed Extended JSON, which is not a lossless round trip for every numeric subtype
Int32 and Double arrive as JSON numbers, so values past JavaScript's safe integer range may be rounded. 64-bit integers keep their exact value
Once relaxed decoding has dropped a subtype or precision it cannot be reconstructed, and generated actions use the decoded value
Canonical Extended JSON import and export are the supported paths when subtype and precision have to round-trip losslessly

Database & Collection Operations

Create databases and collections from the sidebar. Creating a database also creates its first collection, named _init by default
Database names accept letters, numbers, underscores, and hyphens. Collection names also accept dots
Rename, clear (remove all documents), and drop collections. Rename and drop databases
A database rename preserves each collection's documents, indexes, validator, and creation options, and recreates views as views
Copy or duplicate a database or collection, carrying documents and indexes (Pro)
Per-database statistics: collection count, data size, storage size, index size
Per-collection statistics: document count, average document size, data size, storage size, index count and size
Destructive actions require explicit confirmation and honor connection write protection

Licensing & Trial

14-day Pro trial on eligible devices, once, with no payment and no signup
Free capacity: 8 saved connections, 5 favorite queries, 5 pinned collections, and no tab limit
Free authentication: no auth, username/password, SCRAM-SHA-1, and SCRAM-SHA-256
Free soft limits: 500 sampled documents per schema analysis, 500 generated documents, and 250 operation-log entries
Pro removes those limits and adds charts, Cluster Monitor, user and role management, and the MCP server
Pro also adds detached windows, document comparison, custom themes, Excel import/export, NDJSON import, collection copy, and database-level import/export
Pro auth methods: X.509, LDAP, Kerberos, and AWS IAM
Sold monthly, annually, or as a lifetime license, active on up to 3 devices
Devices inactive for 30 days can be removed automatically, and you can deactivate the current one without cancelling
Subscriptions validate at startup, and a network failure does not remove access you already have
An inactive subscription gets a three-day grace period before reverting to Free. A revoked license reverts at once
Lifetime licenses include every future update, with no time limit, and activate on any version
Gated menu items, view buttons, formats, auth choices, and settings identify their Pro requirement
Soft-limited inputs show the Free maximum rather than accepting a value they will reject

Updates, Privacy & Support

Automatic update check on launch with an in-app banner, one-click install and restart, and release notes afterwards
A warning when the current build is no longer eligible for updates
Footer access to settings, updates, the repository, current operations, shortcuts, themes, zoom, logs, MCP, and licensing
Status and settings changes made in one window propagate to the others
Non-blocking notifications for successful, failed, informational, and warning outcomes
A failed optimistic change rolls back to the last confirmed state
Component failures give a recoverable fallback with diagnostic export and a support contact
Anonymous telemetry is opt-out on first launch and from settings, and excludes personally identifiable information
Privacy Policy, Terms of Service, and EULA ship with the product
Three landmark regions for assistive technology: the connection tree, the workspace, and the status bar

Download Monghoul Request a feature

Missing something you need? Tell us on GitHub.

Features

Query Editor

Result Viewer

Charts

Aggregation Builder

Explain View

Connection Management

Inline Editing

Schema Analysis

Cluster Monitoring

Users & Roles

MCP Server (AI Tools)

Import & Export

Index Management

Workspace & Tabs

Sidebar Organization

Data Generation

Operation Logs

Themes

Document Diff

Date Helper

Validation Rules

Complete Feature Reference