In 2022, a quick reply to a data request led to an unexpected idea. The sender was travelling, and the person receiving the data assumed there must be an app that could run MongoDB queries from a phone.
There wasn't. A laptop had been opened and the data forwarded. But the conversation described a useful workflow before the tool existed.
That assumption became the starting point for ClusterLens.
From a quick answer to a complete workflow
The first moment was about speed and convenience. The ambition has since become broader: a full mobile MongoDB client that makes database work practical on an iPhone or iPad.
MongoDB Compass and Atlas set the benchmark for the usefulness and ease the product aims to offer. Bringing comparable breadth to mobile is a development goal. It means thinking beyond a single successful query to the connected tasks that make a database tool useful day after day.
A quick document lookup is one entry point. Understanding a collection, refining a query and working through its results are part of a larger workflow. The product needs to consider how someone moves between those tasks, keeps their place and stays clear about the database they are working with.
For an engineer away from a desk, the intended benefit is continuity: being able to continue useful database work with the device already at hand.
Mobile changes the design
A small screen makes interface decisions more consequential. Connection details, query controls and results all compete for space. Documents can be deeply nested. Results can be large. A network connection can disappear halfway through a task.
The challenge is to preserve capability while making those situations understandable. A broad feature list will only help if the workflows are clear, responsive and reliable on the devices people actually use.
This is where the original conversation remains useful. The imagined app sounded simple because its purpose was easy to understand. Keeping that clarity as the product grows is a design requirement.
Beyond the first page of results
ClusterLens is a developer-stage application built with SwiftUI for iPhone and iPad. Its architecture connects directly to MongoDB over TLS, without a hosted gateway between the app and the database.
The current work adds typed visual filters alongside raw Extended JSON, query templates and projection controls for choosing returned fields. A filter needs to preserve the type of a value, and a readable document needs to preserve its meaning. Treating every input as text would make the interface simpler at the cost of an accurate query.
Paged browsing extends the workflow beyond the first result set. Loading another page should remain a deliberate action, with a clear connection between the query, the current results and the next request. That separation also matters when a request is cancelled or a connection fails. Cursor continuation through getMore still needs focused validation against a real database.
Making exports and writes deliberate
Export is part of that same workflow. The current implementation supports JSON and CSV, selected fields, an explicit row limit and export across result pages. It bounds an export at 10,000 rows or 50 MiB and provides cancellation. Cross-page export is explicitly requested, so browsing a collection does not silently turn into an unbounded download.
CSV output also includes formula protection. A database value can be interpreted as a formula when a spreadsheet opens the file, so the export path has to account for the application that will read it next. These safeguards are part of the implementation; they are not a claim that every spreadsheet or database combination has been tested.
Operations that change the database need a different level of intent from browsing. Create and drop operations are guarded, with the locked state checked in the synthetic workflow. The purpose is to make a consequential action clear before it proceeds. Successful checks of those controls do not establish that real-database writes have been validated.
What the current build establishes
The development checkpoint on 3 October 2026 includes 32 passing simulator XCTest tests, 10 native test groups and a successful unsigned Release build. Manual checks used synthetic data to exercise 35-row exports, cancellation, locked create/drop controls and the iPad layout.
Version 0.3.0 (1) was also installed and launched on an iPhone. The app was then tested on that iPhone with real MongoDB clusters and connection strings, and worked in that setup. This adds a real-cluster check on a physical device to the simulator and synthetic-data checks.
The boundaries of this evidence are useful. Simulator checks can cover parsing, state transitions and controlled failure cases. Synthetic records make exports repeatable. Testing against real clusters establishes that the app works in the tested iPhone setup. It does not establish a complete device QA pass, validation of every database operation or performance under sustained use.
The next proof
Full VoiceOver coverage, compact-phone layouts and physical-device performance still need validation. Cursor continuation through getMore, connection interruptions and real-database writes need focused checks beyond the successful iPhone test. The next work needs to exercise those paths against actual devices and databases, including interruptions and failures.
ClusterLens remains a development build, with no App Store release or production-readiness claim. Full Compass-level workflow breadth and Atlas-style convenience remain the direction, rather than a description of today's coverage.
The 2022 conversation gave the project its starting point. Someone assumed useful MongoDB work could already happen from a phone. Building a capable mobile client around that expectation is the work now.
Imagined by Aekam. Built and curated by Miwl.