Two-step confirmation on drops and deletes is what makes a surface this broad usable against something real, and disabling components lets you cut the tool list down to the shape you trust. Worth knowing: every database it creates gets a metadata collection recording its origin, which you can drop once you have added your own.
A local server with broad coverage of MongoDB: not just find and aggregate, but schema inference, query plans, index management, users, GridFS, transactions, change streams and sharding status.
- Queries with filters, projections and sorting, counts, distinct values, text search, geospatial queries and collation-aware queries
- Aggregation pipelines, execution plan analysis, and live query-pattern analysis that suggests optimisations
- Inferred collection schemas, schema comparison between two collections, and generated JSON Schema validators
- Writes: insert, update, delete, bulk operations, and multi-operation ACID transactions
- Administration: databases and collections, indexes, users, time-series collections, GridFS buckets, validation and sharding status
- Named connection aliases so you can switch between clusters and come back to the original
- Change streams for watching a collection live, and data export as JSON or CSV
Node.js for the npx route, or the published Docker image; source builds are documented too. A MongoDB connection string, supplied at startup or as a named alias. Configuration lives in ~/.mongodb-lens.json and can be overridden by environment variables. Destructive operations - dropping a database, collection, index or user, and deleting documents - require a second confirming call, and individual tools, prompts or resources can be disabled outright.
One command — npx -y mongodb-lens